{ "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "provenance": [] }, "kernelspec": { "name": "python3", "display_name": "Python 3" }, "language_info": { "name": "python" } }, "cells": [ { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "sVL10DfIJoBI", "outputId": "7a697af4-9d4d-4f1a-f5c1-70a43b5b2994" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Mounted at /content/drive/\n" ] } ], "source": [ "# Connecting Google Drive with Google Colab\n", "from google.colab import drive\n", "drive.mount('/content/drive/')" ] }, { "cell_type": "code", "source": [ "# Loading essential libraries\n", "import numpy as np\n", "import pandas as pd\n" ], "metadata": { "id": "8-JaOKsXNPbH" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Loading the dataset\n", "df = pd.read_csv('/content/kaggle_fake_train.csv')" ], "metadata": { "id": "wF2GonnmT6ro" }, "execution_count": null, "outputs": [] }, { "cell_type": "markdown", "source": [ "**Exploring the dataset**" ], "metadata": { "id": "APUrxEpNT_54" } }, { "cell_type": "code", "source": [ "\n", "df.shape" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "HNa6uklFWfzu", "outputId": "3b28d58c-a5bd-48e7-f165-f506edcdfb2b" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(20800, 5)" ] }, "metadata": {}, "execution_count": 17 } ] }, { "cell_type": "code", "source": [ "df.columns" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "ugc5LD_rXDGu", "outputId": "8636ad9c-c517-4392-d522-75321236283a" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "Index(['id', 'title', 'author', 'text', 'label'], dtype='object')" ] }, "metadata": {}, "execution_count": 18 } ] }, { "cell_type": "code", "source": [ "df.head()" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 206 }, "id": "CIdin-C4XGPl", "outputId": "79724189-5df1-4205-cc8f-80ce4e4bbf7f" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ " id title author \\\n", "0 0 House Dem Aide: We Didn’t Even See Comey’s Let... Darrell Lucus \n", "1 1 FLYNN: Hillary Clinton, Big Woman on Campus - ... Daniel J. Flynn \n", "2 2 Why the Truth Might Get You Fired Consortiumnews.com \n", "3 3 15 Civilians Killed In Single US Airstrike Hav... Jessica Purkiss \n", "4 4 Iranian woman jailed for fictional unpublished... Howard Portnoy \n", "\n", " text label \n", "0 House Dem Aide: We Didn’t Even See Comey’s Let... 1 \n", "1 Ever get the feeling your life circles the rou... 0 \n", "2 Why the Truth Might Get You Fired October 29, ... 1 \n", "3 Videos 15 Civilians Killed In Single US Airstr... 1 \n", "4 Print \\nAn Iranian woman has been sentenced to... 1 " ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
idtitleauthortextlabel
00House Dem Aide: We Didn’t Even See Comey’s Let...Darrell LucusHouse Dem Aide: We Didn’t Even See Comey’s Let...1
11FLYNN: Hillary Clinton, Big Woman on Campus - ...Daniel J. FlynnEver get the feeling your life circles the rou...0
22Why the Truth Might Get You FiredConsortiumnews.comWhy the Truth Might Get You Fired October 29, ...1
3315 Civilians Killed In Single US Airstrike Hav...Jessica PurkissVideos 15 Civilians Killed In Single US Airstr...1
44Iranian woman jailed for fictional unpublished...Howard PortnoyPrint \\nAn Iranian woman has been sentenced to...1
\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", "\n", "\n", "\n", " \n", "
\n", "\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "df", "summary": "{\n \"name\": \"df\",\n \"rows\": 20800,\n \"fields\": [\n {\n \"column\": \"id\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 6004,\n \"min\": 0,\n \"max\": 20799,\n \"num_unique_values\": 20800,\n \"samples\": [\n 14649,\n 9231,\n 6473\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"title\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 19803,\n \"samples\": [\n \"Florida Republicans Voting for Hillary Clinton \\u00b7 Guardian Liberty Voice\",\n \"\\u2018Weekend Warriors\\u2019 Show Survival Benefits - The New York Times\",\n \"Part 6, Democratic Network Money: Commons Money that Works for Us, Instead of Us Working for It\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"author\",\n \"properties\": {\n \"dtype\": \"category\",\n \"num_unique_values\": 4201,\n \"samples\": [\n \"AARGH63\",\n \"Dave Kehr\",\n \"Patrick Healy\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"text\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 20386,\n \"samples\": [\n \"Multan a un ultrasur por lanzar a un jugador del Bar\\u00e7a al campo durante el partido Madrid-Legia EL PARTIDO SE CONSIDERABA DE ALTO RIESGO PARA LOS JUGADORES DEL BARCELONA Real Madrid \\nUn jugador barcelonista golpe\\u00f3 a varios jugadores durante el encuentro de la Champions que tuvo lugar ayer entre el Real Madrid y el Legia Varsovia despu\\u00e9s de ser arrojado desde la grada por un hincha madridista. \\nEl portero del Legia, muy molesto, intent\\u00f3 devolver el jugador a las gradas de una patada y se encar\\u00f3 con los Ultrasur, que ya preparaban a otro jugador para echarlo al campo y entorpecer el encuentro. En esta ocasi\\u00f3n pretend\\u00edan prenderlo con un mechero antes de lanzarlo al c\\u00e9sped. \\nLas autoridades han vuelto a pedir a los aficionados que dejen de lanzar cosas al campo como botellas, monedas y jugadores del Barcelona. El dispositivo de seguridad no puedo hacer nada por evitar que algunos hinchas introdujeran a jugadores barcelonistas a la grada, pues los llevaban camuflados dentro de bocadillos de gran tama\\u00f1o. \\nEl partido pudo ser reanudado cuando el \\u00e1rbitro apart\\u00f3 a un rinc\\u00f3n del c\\u00e9sped al jugador barcelonista.\",\n \"Written by Jacob G. Hornberger As US officials continue to accuse Russia of meddling with the US presidential election, an accusation that they have provided no evidence whatsoever to support, let\\u2019s review some of the US government\\u2019s history of meddling with elections in others countries.1. In 1951, the democratically elected parliament of Iran elected a man named Mohammad Mossadegh to be Iran\\u2019s prime minister. Mossadegh angered British Empire officials by nationalizing British oil interests in the country.British officials then turned to the CIA for assistance. In 1953, the CIA secretly fomented a violent coup in Iran, which succeeded in ousting Mossadegh from power and making the Shah of Iran the supreme unelected dictator of the country.To fortify the Shah\\u2019s dictatorial hold on power, the CIA helped organize and train the his domestic police force, the Savak, which was essentially a combination of the CIA, the NSA, and the military. Part of the CIA\\u2019s training involved teaching Savak agents the art of torture.For the next 26 years, the Iranian people suffered under one of the most brutal and tyrannical dictatorships in the world, one that US officials fully supported and called an ally and friend of the United States.In 1979, Iranians successfully revolted against the Shah\\u2019s regime and ousted him from power. One result was not a restoration of the democratic system that had elected Mossadegh but rather another brutal dictatorship, this time a religious one. Another result is the bad relations between the Iran and US governments that continues to exist today.2. In 1951, the Guatemalan people democratically elected a man named Jacobo Arbenz to be their president. Arbenz, however, was not satisfactory to US officials, especially the national-security branch of the government, specifically the Pentagon and the CIA. The reason that US officials opposed Arbenz was that he was a socialist, and US officials considered a socialist president of Guatemala to be a threat to \\u201cnational security\\u201d here in the United States.In 1954 \\u2014 one year after the coup in Iran, the CIA fomented a violent military coup that succeeded in removing Arbenz from power and replacing him with one of the most brutal unelected military dictators in Latin American history, a man named Carlos Castillo Armas. The CIA had a kill list prepared for the coup, which Arbenz was able to escape by fleeing the country before Castillo was able to get him. The CIA\\u2019s destruction of Guatemala\\u2019s democratic system threw the nation into a 30-year civil war that ended up killing millions of Guatemalan people, especially many of the poor.3. In 1960 a man named Patrice Lamumba was elected Congo\\u2019s first prime minister after independence from Belgium. Lamumba spoke out against Western imperialism and refused to take sides in the Cold War, which caused the CIA to conclude that he was a threat to \\u201cnational security.\\u201d The CIA orchestrated the assassination of Lamumba, which ended up taking place on January 17, 1961, just three days before President Kennedy, who liked Lamumba and who would have ordered the CIA to stand down, was to be sworn into office.4. In 1970 a man named Salvador Allende received a plurality of votes in the presidential election in Chile. Pursuant to the Chilean constitution, the election was thrown into the national congress. President Richard Nixon, his national-security team, the Pentagon, and the CIA concluded that because Allende believed in communism and socialism, he posed a grave threat to \\u201cnational security\\u201d here in the United States. The CIA attempted to bribe members of the congress to vote against Allende. It also orchestrated the kidnapping of the head of Chile\\u2019s armed forces, Gen Rene Schneider, who opposed a US military coup in his country, especially since a coup would violate the country\\u2019s constitution. The kidnapping attempt on Schneider left him dead.The CIA then fomented a coup that took place on 9/11 1973 that violently ousted Allende from power and left him dead. Replacing him was army Gen. Augusto Pinochet, one of the most brutal unelected military dictators in history. By the time Pinochet\\u2019s 17-year reign of military terror came to an end in 1990, he and his CIA-supported goons had incarcerated, raped, tortured, or killed tens of thousands of innocent people \\u2014 that is, people whose only \\u201ccrime\\u201d was believing in socialism \\u2014 with the full support of the CIA, Pentagon, Nixon, and his \\u201cnational security\\u201d team.Of course, there are also the more recent support of regime-change operations that ousted democratically elected presidents that the US government disapproved of, such as in Ukraine and Egypt.And then there is the long list of countries where unelected dictators were targeted for regime change by the US national security state and, where successful, replaced with a brutal unelected pro-US dictator. Cuba, Indonesia, Iraq, Libya, and Syria all come to mind.US officials need to keep in mind that when they point their accusatory index finger at Russia for supposedly meddling in the US presidential election, US officials have, at the same time, three fingers pointing back at themselves. Reprinted with permission from the Future of Freedom Foundation . Related\",\n \"When hearing of an establishment called the Black and Blue Steakhouse, one could reasonably surmise that some, or at the very least one, extremely tough person could be found somewhere in there. Well, the three terrorists who attacked London Bridge over the weekend found the Black and Blue Steakhouse, and they found the toughest man in the house. [Or, maybe more to the point, he found them. Roy Larner, 47, is a huge fan of the South East London soccer team Millwall F. C. The team is known as the Lions, and when terrorists entered the Black and Blue to kill innocent people, Larner let them know who is the true king of the jungle. After the terrorists entered, with obvious bad intentions, Larner sprang into action: \\u201cLike an idiot I shouted back at them. I thought, \\u2018I need to take the p*** out of these b*****s.\\u2019 \\u201cI took a few steps towards them and said, \\u2018F*** you, I\\u2019m Millwall.\\u2019 So they started attacking me. \\u201cI stood in front of them trying to fight them off. Everyone else ran to the back. I was on my own against all three of them, that\\u2019s why I got hurt so much. It was just me, trying to grab them with my bare hands and hold on. I was swinging. \\u201cI got stabbed and sliced eight times. They got me in my head, chest and both hands. There was blood everywhere. \\u201cThey were saying, \\u2018Islam, Islam!\\u2019 I said again, \\u2018F*** you, I\\u2019m Millwall!\\u2019 \\u201cIt was the worst thing I could have done as they carried on attacking me. \\u201cLuckily, none of the blows were straight at me or I\\u2019d be dead. \\u201d A petition has begun, with the aim of awarding Larner with the George Cross for bravery. The award is the second highest commendation granted by Great Britain. Looking at his scars, it\\u2019s really hard to imagine that he won\\u2019t get it. If Millwall is looking for a new slogan, they may have found it. \\u201cF*** you, I\\u2019m Millwall!\\u201d has a nice ring to it. Follow Dylan Gwinn on Twitter: @themightygwinn\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"label\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 0,\n \"min\": 0,\n \"max\": 1,\n \"num_unique_values\": 2,\n \"samples\": [\n 0,\n 1\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}" } }, "metadata": {}, "execution_count": 19 } ] }, { "cell_type": "code", "source": [ "# Dropping the 'id' column\n", "df.drop('id', axis=1, inplace=True)" ], "metadata": { "id": "N2E8Vm6iXK5N" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "df.columns\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "DHHs0TF2XNlm", "outputId": "5e083f47-d3b3-40c4-9407-c38a7c480243" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "Index(['title', 'author', 'text', 'label'], dtype='object')" ] }, "metadata": {}, "execution_count": 21 } ] }, { "cell_type": "code", "source": [ "# Importing essential libraries for visualization\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "import pandas as pd\n", "%matplotlib inline\n", "\n", "df = pd.read_csv('/content/kaggle_fake_train.csv')\n", "\n", "# Visualizing the count of 'fake news' in the dataset\n", "plt.figure(figsize=(10,7))\n", "sns.countplot(x='label', data=df)\n", "plt.xlabel('News Classification')\n", "plt.ylabel('Count')" ], "metadata": { "id": "hTyI1GEYXQqd", "colab": { "base_uri": "https://localhost:8080/", "height": 637 }, "outputId": "7ab056b8-c10b-4a2a-a9a5-193a69a97598" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "Text(0, 0.5, 'Count')" ] }, "metadata": {}, "execution_count": 8 }, { "output_type": "display_data", "data": { "text/plain": [ "
" ], "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2QAAAJaCAYAAABA74ShAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA0VklEQVR4nO3de7TVdb3v/9fijpe18JJrwY6UygsoeQFDMk1zHZdKbjlSR4uUisTaYCnlhRPipQuKd5QTaVvRsXVnnpFu00IRRUqWqCheEM08lO4MsBBWsBUU5u+PNvPnCrRA9LOEx2OMOQbz+/3MOd/f1RjN8fQ753fWVCqVSgAAAHjPtSs9AAAAwJZKkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUEiH0gNsLtasWZOXXnop2267bWpqakqPAwAAFFKpVPKXv/wlPXr0SLt2b38OTJBtIi+99FJ69uxZegwAAKCNePHFF/PBD37wbdcIsk1k2223TfLXP3ptbW3haQAAgFJaWlrSs2fPaiO8HUG2iaz9mGJtba0gAwAA/qGvMrmoBwAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAArpUHoAAOC90e/0G0qPALBJzbnoxNIjvGPOkAEAABTiDNkWwn8VBTY3m8N/FQUAZ8gAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUEjRIJs5c2aOPvro9OjRIzU1Nbntttta7a9UKhk3bly6d++erl27prGxMc8991yrNUuWLMnQoUNTW1ubbt26Zfjw4Vm+fHmrNU888UQOOuigdOnSJT179syECRPWmeWWW27JHnvskS5duqRv3775xS9+scmPFwAA4M2KBtmKFSuy9957Z9KkSevdP2HChEycODGTJ0/O7Nmzs/XWW6epqSmvvfZadc3QoUMzb968TJs2LXfccUdmzpyZESNGVPe3tLTk8MMPz84775w5c+bkoosuyrnnnpurr766umbWrFn5/Oc/n+HDh+exxx7L4MGDM3jw4Dz11FPv3sEDAABbvJpKpVIpPUSS1NTU5NZbb83gwYOT/PXsWI8ePfKtb30r3/72t5Mky5YtS319faZMmZLjjz8+8+fPT58+ffLwww+nf//+SZKpU6fmqKOOyn/+53+mR48e+eEPf5jvfOc7WbhwYTp16pQkOeuss3LbbbflmWeeSZIcd9xxWbFiRe64447qPAcccED22WefTJ48+R+av6WlJXV1dVm2bFlqa2s31Z9lk+l3+g2lRwDYpOZcdGLpEd53vBcAm5u2+l6wIW3QZr9DtmDBgixcuDCNjY3VbXV1dRkwYECam5uTJM3NzenWrVs1xpKksbEx7dq1y+zZs6trDj744GqMJUlTU1OeffbZvPLKK9U1b36dtWvWvs76rFy5Mi0tLa1uAAAAG6LNBtnChQuTJPX19a2219fXV/ctXLgwO+20U6v9HTp0yPbbb99qzfqe482v8VZr1u5fn/Hjx6eurq5669mz54YeIgAAsIVrs0HW1o0ZMybLli2r3l588cXSIwEAAO8zbTbIGhoakiSLFi1qtX3RokXVfQ0NDVm8eHGr/W+88UaWLFnSas36nuPNr/FWa9buX5/OnTuntra21Q0AAGBDtNkg69WrVxoaGjJ9+vTqtpaWlsyePTsDBw5MkgwcODBLly7NnDlzqmvuvfferFmzJgMGDKiumTlzZl5//fXqmmnTpmX33XfPdtttV13z5tdZu2bt6wAAALwbigbZ8uXLM3fu3MydOzfJXy/kMXfu3LzwwgupqanJqaeemu9973u5/fbb8+STT+bEE09Mjx49qldi7N27d4444oicdNJJeeihh/LAAw9k1KhROf7449OjR48kyRe+8IV06tQpw4cPz7x583LzzTfniiuuyOjRo6tzfPOb38zUqVNzySWX5Jlnnsm5556bRx55JKNGjXqv/yQAAMAWpEPJF3/kkUdy6KGHVu+vjaRhw4ZlypQpOeOMM7JixYqMGDEiS5cuzSc/+clMnTo1Xbp0qT7mxhtvzKhRo3LYYYelXbt2GTJkSCZOnFjdX1dXl7vvvjsjR45Mv379suOOO2bcuHGtfqvsE5/4RG666aaMHTs2//t//+/suuuuue2227LXXnu9B38FAABgS9Vmfofs/c7vkAG8t9rqb8+0Zd4LgM1NW30v2Cx+hwwAAGBzJ8gAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUEibDrLVq1fn7LPPTq9evdK1a9d85CMfyXe/+91UKpXqmkqlknHjxqV79+7p2rVrGhsb89xzz7V6niVLlmTo0KGpra1Nt27dMnz48CxfvrzVmieeeCIHHXRQunTpkp49e2bChAnvyTECAABbrjYdZBdeeGF++MMf5qqrrsr8+fNz4YUXZsKECbnyyiurayZMmJCJEydm8uTJmT17drbeeus0NTXltddeq64ZOnRo5s2bl2nTpuWOO+7IzJkzM2LEiOr+lpaWHH744dl5550zZ86cXHTRRTn33HNz9dVXv6fHCwAAbFk6lB7g7cyaNSvHHHNMBg0alCTZZZdd8u///u956KGHkvz17Njll1+esWPH5phjjkmS3HDDDamvr89tt92W448/PvPnz8/UqVPz8MMPp3///kmSK6+8MkcddVQuvvji9OjRIzfeeGNWrVqVa6+9Np06dcqee+6ZuXPn5tJLL20VbgAAAJtSmz5D9olPfCLTp0/Pb37zmyTJ448/nl//+tc58sgjkyQLFizIwoUL09jYWH1MXV1dBgwYkObm5iRJc3NzunXrVo2xJGlsbEy7du0ye/bs6pqDDz44nTp1qq5pamrKs88+m1deeWW9s61cuTItLS2tbgAAABuiTZ8hO+uss9LS0pI99tgj7du3z+rVq/P9738/Q4cOTZIsXLgwSVJfX9/qcfX19dV9CxcuzE477dRqf4cOHbL99tu3WtOrV691nmPtvu22226d2caPH5/zzjtvExwlAACwpWrTZ8h++tOf5sYbb8xNN92URx99NNdff30uvvjiXH/99aVHy5gxY7Js2bLq7cUXXyw9EgAA8D7Tps+QnX766TnrrLNy/PHHJ0n69u2b3//+9xk/fnyGDRuWhoaGJMmiRYvSvXv36uMWLVqUffbZJ0nS0NCQxYsXt3reN954I0uWLKk+vqGhIYsWLWq1Zu39tWv+VufOndO5c+d3fpAAAMAWq02fIfuv//qvtGvXesT27dtnzZo1SZJevXqloaEh06dPr+5vaWnJ7NmzM3DgwCTJwIEDs3Tp0syZM6e65t57782aNWsyYMCA6pqZM2fm9ddfr66ZNm1adt999/V+XBEAAGBTaNNBdvTRR+f73/9+7rzzzvzud7/LrbfemksvvTT/83/+zyRJTU1NTj311Hzve9/L7bffnieffDInnnhievTokcGDBydJevfunSOOOCInnXRSHnrooTzwwAMZNWpUjj/++PTo0SNJ8oUvfCGdOnXK8OHDM2/evNx888254oorMnr06FKHDgAAbAHa9EcWr7zyypx99tn5l3/5lyxevDg9evTIySefnHHjxlXXnHHGGVmxYkVGjBiRpUuX5pOf/GSmTp2aLl26VNfceOONGTVqVA477LC0a9cuQ4YMycSJE6v76+rqcvfdd2fkyJHp169fdtxxx4wbN84l7wEAgHdVTaVSqZQeYnPQ0tKSurq6LFu2LLW1taXHWUe/028oPQLAJjXnohNLj/C+470A2Ny01feCDWmDNv2RRQAAgM2ZIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgkDYfZH/4wx/yxS9+MTvssEO6du2avn375pFHHqnur1QqGTduXLp3756uXbumsbExzz33XKvnWLJkSYYOHZra2tp069Ytw4cPz/Lly1uteeKJJ3LQQQelS5cu6dmzZyZMmPCeHB8AALDlatNB9sorr+TAAw9Mx44d88tf/jJPP/10Lrnkkmy33XbVNRMmTMjEiRMzefLkzJ49O1tvvXWampry2muvVdcMHTo08+bNy7Rp03LHHXdk5syZGTFiRHV/S0tLDj/88Oy8886ZM2dOLrroopx77rm5+uqr39PjBQAAtiwdSg/wdi688ML07Nkz1113XXVbr169qv+uVCq5/PLLM3bs2BxzzDFJkhtuuCH19fW57bbbcvzxx2f+/PmZOnVqHn744fTv3z9JcuWVV+aoo47KxRdfnB49euTGG2/MqlWrcu2116ZTp07Zc889M3fu3Fx66aWtwg0AAGBTatNnyG6//fb0798/n/vc57LTTjtl3333zTXXXFPdv2DBgixcuDCNjY3VbXV1dRkwYECam5uTJM3NzenWrVs1xpKksbEx7dq1y+zZs6trDj744HTq1Km6pqmpKc8++2xeeeWV9c62cuXKtLS0tLoBAABsiDYdZP/v//2//PCHP8yuu+6au+66K1//+tfzjW98I9dff32SZOHChUmS+vr6Vo+rr6+v7lu4cGF22mmnVvs7dOiQ7bffvtWa9T3Hm1/jb40fPz51dXXVW8+ePd/h0QIAAFuaNh1ka9asyX777Zcf/OAH2XfffTNixIicdNJJmTx5cunRMmbMmCxbtqx6e/HFF0uPBAAAvM+06SDr3r17+vTp02pb796988ILLyRJGhoakiSLFi1qtWbRokXVfQ0NDVm8eHGr/W+88UaWLFnSas36nuPNr/G3OnfunNra2lY3AACADdGmg+zAAw/Ms88+22rbb37zm+y8885J/nqBj4aGhkyfPr26v6WlJbNnz87AgQOTJAMHDszSpUszZ86c6pp77703a9asyYABA6prZs6cmddff726Ztq0adl9991bXdERAABgU2rTQXbaaaflwQcfzA9+8IP89re/zU033ZSrr746I0eOTJLU1NTk1FNPzfe+973cfvvtefLJJ3PiiSemR48eGTx4cJK/nlE74ogjctJJJ+Whhx7KAw88kFGjRuX4449Pjx49kiRf+MIX0qlTpwwfPjzz5s3LzTffnCuuuCKjR48udegAAMAWoE1f9n7//ffPrbfemjFjxuT8889Pr169cvnll2fo0KHVNWeccUZWrFiRESNGZOnSpfnkJz+ZqVOnpkuXLtU1N954Y0aNGpXDDjss7dq1y5AhQzJx4sTq/rq6utx9990ZOXJk+vXrlx133DHjxo1zyXsAAOBdVVOpVCqlh9gctLS0pK6uLsuWLWuT3yfrd/oNpUcA2KTmXHRi6RHed7wXAJubtvpesCFt0KY/sggAALA526gg+/CHP5w///nP62xfunRpPvzhD7/joQAAALYEGxVkv/vd77J69ep1tq9cuTJ/+MMf3vFQAAAAW4INuqjH7bffXv33XXfdlbq6uur91atXZ/r06dlll1022XAAAACbsw0KsrWXkq+pqcmwYcNa7evYsWN22WWXXHLJJZtsOAAAgM3ZBgXZmjVrkvz1B5kffvjh7Ljjju/KUAAAAFuCjfodsgULFmzqOQAAALY4G/3D0NOnT8/06dOzePHi6pmzta699tp3PBgAAMDmbqOC7Lzzzsv555+f/v37p3v37qmpqdnUcwEAAGz2NirIJk+enClTpuSEE07Y1PMAAABsMTbqd8hWrVqVT3ziE5t6FgAAgC3KRgXZV7/61dx0002behYAAIAtykZ9ZPG1117L1VdfnXvuuScf+9jH0rFjx1b7L7300k0yHAAAwOZso4LsiSeeyD777JMkeeqpp1rtc4EPAACAf8xGBdl99923qecAAADY4mzUd8gAAAB45zbqDNmhhx76th9NvPfeezd6IAAAgC3FRgXZ2u+PrfX6669n7ty5eeqppzJs2LBNMRcAAMBmb6OC7LLLLlvv9nPPPTfLly9/RwMBAABsKTbpd8i++MUv5tprr92UTwkAALDZ2qRB1tzcnC5dumzKpwQAANhsbdRHFo899thW9yuVSv74xz/mkUceydlnn71JBgMAANjcbVSQ1dXVtbrfrl277L777jn//PNz+OGHb5LBAAAANncbFWTXXXfdpp4DAABgi7NRQbbWnDlzMn/+/CTJnnvumX333XeTDAUAALAl2KggW7x4cY4//vjMmDEj3bp1S5IsXbo0hx56aH7yk5/kAx/4wKacEQAAYLO0UVdZPOWUU/KXv/wl8+bNy5IlS7JkyZI89dRTaWlpyTe+8Y1NPSMAAMBmaaPOkE2dOjX33HNPevfuXd3Wp0+fTJo0yUU9AAAA/kEbdYZszZo16dix4zrbO3bsmDVr1rzjoQAAALYEGxVkn/70p/PNb34zL730UnXbH/7wh5x22mk57LDDNtlwAAAAm7ONCrKrrroqLS0t2WWXXfKRj3wkH/nIR9KrV6+0tLTkyiuv3NQzAgAAbJY26jtkPXv2zKOPPpp77rknzzzzTJKkd+/eaWxs3KTDAQAAbM426AzZvffemz59+qSlpSU1NTX5H//jf+SUU07JKaeckv333z977rlnfvWrX71bswIAAGxWNijILr/88px00kmpra1dZ19dXV1OPvnkXHrppZtsOAAAgM3ZBgXZ448/niOOOOIt9x9++OGZM2fOOx4KAABgS7BBQbZo0aL1Xu5+rQ4dOuTll19+x0MBAABsCTYoyP7pn/4pTz311Fvuf+KJJ9K9e/d3PBQAAMCWYIOC7KijjsrZZ5+d1157bZ19r776as4555x85jOf2WTDAQAAbM426LL3Y8eOzc9+9rPstttuGTVqVHbfffckyTPPPJNJkyZl9erV+c53vvOuDAoAALC52aAgq6+vz6xZs/L1r389Y8aMSaVSSZLU1NSkqakpkyZNSn19/bsyKAAAwOZmg38Yeuedd84vfvGLvPLKK/ntb3+bSqWSXXfdNdttt927MR8AAMBma4ODbK3tttsu+++//6acBQAAYIuyQRf1AAAAYNMRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKCQ91WQXXDBBampqcmpp55a3fbaa69l5MiR2WGHHbLNNttkyJAhWbRoUavHvfDCCxk0aFC22mqr7LTTTjn99NPzxhtvtFozY8aM7LfffuncuXM++tGPZsqUKe/BEQEAAFuy902QPfzww/nRj36Uj33sY622n3baafn5z3+eW265Jffff39eeumlHHvssdX9q1evzqBBg7Jq1arMmjUr119/faZMmZJx48ZV1yxYsCCDBg3KoYcemrlz5+bUU0/NV7/61dx1113v2fEBAABbnvdFkC1fvjxDhw7NNddck+222666fdmyZfnXf/3XXHrppfn0pz+dfv365brrrsusWbPy4IMPJknuvvvuPP300/m3f/u37LPPPjnyyCPz3e9+N5MmTcqqVauSJJMnT06vXr1yySWXpHfv3hk1alQ++9nP5rLLLityvAAAwJbhfRFkI0eOzKBBg9LY2Nhq+5w5c/L666+32r7HHnvkQx/6UJqbm5Mkzc3N6du3b+rr66trmpqa0tLSknnz5lXX/O1zNzU1VZ8DAADg3dCh9AB/z09+8pM8+uijefjhh9fZt3DhwnTq1CndunVrtb2+vj4LFy6srnlzjK3dv3bf261paWnJq6++mq5du67z2itXrszKlSur91taWjb84AAAgC1amz5D9uKLL+ab3/xmbrzxxnTp0qX0OK2MHz8+dXV11VvPnj1LjwQAALzPtOkgmzNnThYvXpz99tsvHTp0SIcOHXL//fdn4sSJ6dChQ+rr67Nq1aosXbq01eMWLVqUhoaGJElDQ8M6V11ce//vramtrV3v2bEkGTNmTJYtW1a9vfjii5vikAEAgC1Imw6yww47LE8++WTmzp1bvfXv3z9Dhw6t/rtjx46ZPn169THPPvtsXnjhhQwcODBJMnDgwDz55JNZvHhxdc20adNSW1ubPn36VNe8+TnWrln7HOvTuXPn1NbWtroBAABsiDb9HbJtt902e+21V6ttW2+9dXbYYYfq9uHDh2f06NHZfvvtU1tbm1NOOSUDBw7MAQcckCQ5/PDD06dPn5xwwgmZMGFCFi5cmLFjx2bkyJHp3LlzkuRrX/tarrrqqpxxxhn5yle+knvvvTc//elPc+edd763BwwAAGxR2nSQ/SMuu+yytGvXLkOGDMnKlSvT1NSU//N//k91f/v27XPHHXfk61//egYOHJitt946w4YNy/nnn19d06tXr9x555057bTTcsUVV+SDH/xgfvzjH6epqanEIQEAAFuImkqlUik9xOagpaUldXV1WbZsWZv8+GK/028oPQLAJjXnohNLj/C+470A2Ny01feCDWmDNv0dMgAAgM2ZIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgkDYdZOPHj8/++++fbbfdNjvttFMGDx6cZ599ttWa1157LSNHjswOO+yQbbbZJkOGDMmiRYtarXnhhRcyaNCgbLXVVtlpp51y+umn54033mi1ZsaMGdlvv/3SuXPnfPSjH82UKVPe7cMDAAC2cG06yO6///6MHDkyDz74YKZNm5bXX389hx9+eFasWFFdc9ppp+XnP/95brnlltx///156aWXcuyxx1b3r169OoMGDcqqVasya9asXH/99ZkyZUrGjRtXXbNgwYIMGjQohx56aObOnZtTTz01X/3qV3PXXXe9p8cLAABsWWoqlUql9BD/qJdffjk77bRT7r///hx88MFZtmxZPvCBD+Smm27KZz/72STJM888k969e6e5uTkHHHBAfvnLX+Yzn/lMXnrppdTX1ydJJk+enDPPPDMvv/xyOnXqlDPPPDN33nlnnnrqqeprHX/88Vm6dGmmTp36D83W0tKSurq6LFu2LLW1tZv+4N+hfqffUHoEgE1qzkUnlh7hfcd7AbC5aavvBRvSBm36DNnfWrZsWZJk++23T5LMmTMnr7/+ehobG6tr9thjj3zoQx9Kc3NzkqS5uTl9+/atxliSNDU1paWlJfPmzauuefNzrF2z9jnWZ+XKlWlpaWl1AwAA2BDvmyBbs2ZNTj311Bx44IHZa6+9kiQLFy5Mp06d0q1bt1Zr6+vrs3DhwuqaN8fY2v1r973dmpaWlrz66qvrnWf8+PGpq6ur3nr27PmOjxEAANiyvG+CbOTIkXnqqafyk5/8pPQoSZIxY8Zk2bJl1duLL75YeiQAAOB9pkPpAf4Ro0aNyh133JGZM2fmgx/8YHV7Q0NDVq1alaVLl7Y6S7Zo0aI0NDRU1zz00EOtnm/tVRjfvOZvr8y4aNGi1NbWpmvXruudqXPnzuncufM7PjYAAGDL1abPkFUqlYwaNSq33npr7r333vTq1avV/n79+qVjx46ZPn16dduzzz6bF154IQMHDkySDBw4ME8++WQWL15cXTNt2rTU1tamT58+1TVvfo61a9Y+BwAAwLuhTZ8hGzlyZG666ab8x3/8R7bddtvqd77q6urStWvX1NXVZfjw4Rk9enS233771NbW5pRTTsnAgQNzwAEHJEkOP/zw9OnTJyeccEImTJiQhQsXZuzYsRk5cmT1DNfXvva1XHXVVTnjjDPyla98Jffee29++tOf5s477yx27AAAwOavTZ8h++EPf5hly5blkEMOSffu3au3m2++ubrmsssuy2c+85kMGTIkBx98cBoaGvKzn/2sur99+/a544470r59+wwcODBf/OIXc+KJJ+b888+vrunVq1fuvPPOTJs2LXvvvXcuueSS/PjHP05TU9N7erwAAMCWpU2fIftHfiKtS5cumTRpUiZNmvSWa3beeef84he/eNvnOeSQQ/LYY49t8IwAAAAbq02fIQMAANicCTIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLIAAAAChFkAAAAhQgyAACAQgQZAABAIYIMAACgEEEGAABQiCADAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIUIMgAAgEIEGQAAQCGCDAAAoBBBBgAAUIggAwAAKESQAQAAFCLI/sakSZOyyy67pEuXLhkwYEAeeuih0iMBAACbKUH2JjfffHNGjx6dc845J48++mj23nvvNDU1ZfHixaVHAwAANkOC7E0uvfTSnHTSSfnyl7+cPn36ZPLkydlqq61y7bXXlh4NAADYDAmy/7Zq1arMmTMnjY2N1W3t2rVLY2NjmpubC04GAABsrjqUHqCt+NOf/pTVq1envr6+1fb6+vo888wz66xfuXJlVq5cWb2/bNmyJElLS8u7O+hGWr3y1dIjAGxSbfX/b9sy7wXA5qatvhesnatSqfzdtYJsI40fPz7nnXfeOtt79uxZYBqALU/dlV8rPQIAhbX194K//OUvqaure9s1guy/7bjjjmnfvn0WLVrUavuiRYvS0NCwzvoxY8Zk9OjR1ftr1qzJkiVLssMOO6SmpuZdnxfaopaWlvTs2TMvvvhiamtrS48DQAHeC+CvZ8b+8pe/pEePHn93rSD7b506dUq/fv0yffr0DB48OMlfI2v69OkZNWrUOus7d+6czp07t9rWrVu392BSaPtqa2u9CQNs4bwXsKX7e2fG1hJkbzJ69OgMGzYs/fv3z8c//vFcfvnlWbFiRb785S+XHg0AANgMCbI3Oe644/Lyyy9n3LhxWbhwYfbZZ59MnTp1nQt9AAAAbAqC7G+MGjVqvR9RBP6+zp0755xzzlnn47wAbDm8F8CGqan8I9diBAAAYJPzw9AAAACFCDIAAIBCBBkAAEAhggwAAKAQQQZsEpMmTcouu+ySLl26ZMCAAXnooYdKjwTAe2zmzJk5+uij06NHj9TU1OS2224rPRK0eYIMeMduvvnmjB49Ouecc04effTR7L333mlqasrixYtLjwbAe2jFihXZe++9M2nSpNKjwPuGy94D79iAAQOy//7756qrrkqSrFmzJj179swpp5ySs846q/B0AJRQU1OTW2+9NYMHDy49CrRpzpAB78iqVasyZ86cNDY2Vre1a9cujY2NaW5uLjgZAEDbJ8iAd+RPf/pTVq9enfr6+lbb6+vrs3DhwkJTAQC8PwgyAACAQgQZ8I7suOOOad++fRYtWtRq+6JFi9LQ0FBoKgCA9wdBBrwjnTp1Sr9+/TJ9+vTqtjVr1mT69OkZOHBgwckAANq+DqUHAN7/Ro8enWHDhqV///75+Mc/nssvvzwrVqzIl7/85dKjAfAeWr58eX77299W7y9YsCBz587N9ttvnw996EMFJ4O2y2XvgU3iqquuykUXXZSFCxdmn332ycSJEzNgwIDSYwHwHpoxY0YOPfTQdbYPGzYsU6ZMee8HgvcBQQYAAFCI75ABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAOA9ZgxY0ZqamqydOnSd/21zj333Oyzzz7rbKuvr09NTU1uu+22fOlLX8rgwYPf9VnWvh4A7w1BBsAm86UvfSk1NTW54IILWm2/7bbbUlNTU2iq9Xvsscfyuc99LvX19enSpUt23XXXnHTSSfnNb37zns/y7W9/O9OnT6/enz9/fs4777z86Ec/yh//+McceeSRueKKKzJlypRN9prri8Ak1dcD4L0hyADYpLp06ZILL7wwr7zySulR3tIdd9yRAw44ICtXrsyNN96Y+fPn59/+7d9SV1eXs88++z2fZ5tttskOO+xQvf/8888nSY455pg0NDSkc+fOqaurS7du3d71Wda+HgDvDUEGwCbV2NiYhoaGjB8//m3X/frXv85BBx2Url27pmfPnvnGN76RFStWJEmuuuqq7LXXXtW1a8+wTZ48udXrjB07Nkny+OOP59BDD822226b2tra9OvXL4888sh6X/e//uu/8uUvfzlHHXVUbr/99jQ2NqZXr14ZMGBALr744vzoRz9a7+P+/Oc/5/Of/3z+6Z/+KVtttVX69u2bf//3f2+15v/+3/+bvn37pmvXrtlhhx3S2NhYPaYZM2bk4x//eLbeeut069YtBx54YH7/+98naX226txzz83RRx+dJGnXrl31zOLffmRxzZo1mTBhQj760Y+mc+fO+dCHPpTvf//71f1nnnlmdtttt2y11Vb58Ic/nLPPPjuvv/56kmTKlCk577zz8vjjj6empiY1NTXVs29/+5HFJ598Mp/+9KerxzRixIgsX768un/tXBdffHG6d++eHXbYISNHjqy+FgBvT5ABsEm1b98+P/jBD3LllVfmP//zP9e75vnnn88RRxyRIUOG5IknnsjNN9+cX//61xk1alSS5FOf+lSefvrpvPzyy0mS+++/PzvuuGNmzJiRJHn99dfT3NycQw45JEkydOjQfPCDH8zDDz+cOXPm5KyzzkrHjh3X+9p33XVX/vSnP+WMM85Y7/63Ogv12muvpV+/frnzzjvz1FNPZcSIETnhhBPy0EMPJfnrR/0+//nP5ytf+Urmz5+fGTNm5Nhjj02lUskbb7yRwYMH51Of+lSeeOKJNDc3Z8SIEev9GOe3v/3tXHfdddXn/OMf/7jeecaMGZMLLrggZ599dp5++uncdNNNqa+vr+7fdtttM2XKlDz99NO54oorcs011+Syyy5Lkhx33HH51re+lT333LP6Gscdd9w6r7FixYo0NTVlu+22y8MPP5xbbrkl99xzT/V/p7Xuu+++PP/887nvvvty/fXXZ8qUKZv045UAm7UKAGwiw4YNqxxzzDGVSqVSOeCAAypf+cpXKpVKpXLrrbdW3vyWM3z48MqIESNaPfZXv/pVpV27dpVXX321smbNmsoOO+xQueWWWyqVSqWyzz77VMaPH19paGioVCqVyq9//etKx44dKytWrKhUKpXKtttuW5kyZco/NOOFF15YSVJZsmTJ26677777Kkkqr7zyyluuGTRoUOVb3/pWpVKpVObMmVNJUvnd7363zro///nPlSSVGTNmrPd5zjnnnMree+9dvf+3f69KpfXftqWlpdK5c+fKNddc87bH8GYXXXRRpV+/fm/5mmslqdx6662VSqVSufrqqyvbbbddZfny5dX9d955Z6Vdu3aVhQsXVufaeeedK2+88UZ1zec+97nKcccd9w/PBrAlc4YMgHfFhRdemOuvvz7z589fZ9/jjz+eKVOmZJtttqnempqasmbNmixYsCA1NTU5+OCDM2PGjCxdujRPP/10/uVf/iUrV67MM888k/vvvz/7779/ttpqqyTJ6NGj89WvfjWNjY254IILqt/BWp9KpbJRx7N69ep897vfTd++fbP99ttnm222yV133ZUXXnghSbL33nvnsMMOS9++ffO5z30u11xzTfV7dNtvv32+9KUvpampKUcffXSuuOKKtzzz9Y+YP39+Vq5cmcMOO+wt19x888058MAD09DQkG222SZjx46tzrohr7P33ntn6623rm478MADs2bNmjz77LPVbXvuuWfat29fvd+9e/csXrx4g14LYEslyAB4Vxx88MFpamrKmDFj1tm3fPnynHzyyZk7d2719vjjj+e5557LRz7ykSTJIYcckhkzZuRXv/pV9t1339TW1lYj7f7778+nPvWp6vOde+65mTdvXgYNGpR77703ffr0ya233rreuXbbbbckyTPPPLNBx3PRRRfliiuuyJlnnpn77rsvc+fOTVNTU1atWpXkrx/VnDZtWn75y1+mT58+ufLKK7P77rtnwYIFSZLrrrsuzc3N+cQnPpGbb745u+22Wx588MENmmGtrl27vu3+5ubmDB06NEcddVTuuOOOPPbYY/nOd75TnXVT+9uPh9bU1GTNmjXvymsBbG4EGQDvmgsuuCA///nP09zc3Gr7fvvtl6effjof/ehH17l16tQpyf//PbJbbrml+l2xQw45JPfcc08eeOCB6ra1dtttt5x22mm5++67c+yxx1a/h/W3Dj/88Oy4446ZMGHCeve/1e+OPfDAAznmmGPyxS9+MXvvvXc+/OEPr3OJ/Jqamhx44IE577zz8thjj6VTp06twnDffffNmDFjMmvWrOy111656aab3upP97Z23XXXdO3atdWl8t9s1qxZ2XnnnfOd73wn/fv3z6677lq9gMhanTp1yurVq9/2dXr37p3HH3+8emGS5K9/h3bt2mX33XffqNkBaE2QAfCu6du3b4YOHZqJEye22n7mmWdm1qxZGTVqVObOnZvnnnsu//Ef/9HqYhEf+9jHst122+Wmm25qFWS33XZbVq5cmQMPPDBJ8uqrr2bUqFGZMWNGfv/73+eBBx7Iww8/nN69e693pq233jo//vGPc+edd+af//mfc8899+R3v/tdHnnkkZxxxhn52te+tt7H7brrrpk2bVpmzZqV+fPn5+STT86iRYuq+2fPnp0f/OAHeeSRR/LCCy/kZz/7WV5++eX07t07CxYsyJgxY9Lc3Jzf//73ufvuu/Pcc8+95Yx/T5cuXXLmmWfmjDPOyA033JDnn38+Dz74YP71X/+1OusLL7yQn/zkJ3n++eczceLEdc4Y7rLLLlmwYEHmzp2bP/3pT1m5cuU6rzN06NB06dIlw4YNy1NPPZX77rsvp5xySk444YRWFxABYOMJMgDeVeeff/46H1/72Mc+lvvvvz+/+c1vctBBB2XffffNuHHj0qNHj+qampqaHHTQQampqcknP/nJ6uNqa2vTv3//6vea2rdvnz//+c858cQTs9tuu+V//a//lSOPPDLnnXfeW850zDHHZNasWenYsWO+8IUvZI899sjnP//5LFu2LN/73vfW+5ixY8dmv/32S1NTUw455JA0NDS0ugx9bW1tZs6cmaOOOiq77bZbxo4dm0suuSRHHnlkttpqqzzzzDMZMmRIdtttt4wYMSIjR47MySefvLF/1px99tn51re+lXHjxqV379457rjjqt/b+ud//uecdtppGTVqVPbZZ5/MmjVrnd9XGzJkSI444ogceuih+cAHPrDOJfyTZKuttspdd92VJUuWZP/9989nP/vZHHbYYbnqqqs2em4AWqupbOy3mwEAAHhHnCEDAAAoRJABAAAUIsgAAAAKEWQAAACFCDIAAIBCBBkAAEAhggwAAKAQQQYAAFCIIAMAAChEkAEAABQiyAAAAAoRZAAAAIX8f3G6GEg7OQE0AAAAAElFTkSuQmCC\n" }, "metadata": {} } ] }, { "cell_type": "markdown", "source": [ "**Data Cleaning and Preprocessing**" ], "metadata": { "id": "8e8NAH-EhqsE" } }, { "cell_type": "code", "source": [ "print(df.shape)" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "Xq_s3Vbfhvuq", "outputId": "9c7fdfd5-c8aa-4b39-8986-58131726e359" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "(20800, 5)\n" ] } ] }, { "cell_type": "code", "source": [ "# Finding any NaN values\n", "df.isna().any()" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 241 }, "id": "QN83k88niKmS", "outputId": "aee02d35-e8e6-4f6f-dc3f-16acf893f1e1" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "id False\n", "title True\n", "author True\n", "text True\n", "label False\n", "dtype: bool" ], "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
0
idFalse
titleTrue
authorTrue
textTrue
labelFalse
\n", "

" ] }, "metadata": {}, "execution_count": 10 } ] }, { "cell_type": "code", "source": [ "# Dropping NaN values\n", "df.dropna(inplace=True)\n", "print(df.shape)" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "K64gL2DXiRZx", "outputId": "615889fd-36e5-4dc1-fceb-22d3e8db5d5b" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "(18285, 5)\n" ] } ] }, { "cell_type": "code", "source": [ "news = df.copy()" ], "metadata": { "id": "JFiYH2vuiV0S" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "news.reset_index(inplace=True)" ], "metadata": { "id": "TXfVPFHViZYh" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Importing essential libraries for performing Natural Language Processing on 'kaggle_fake_train' dataset\n", "import nltk\n", "import re\n", "nltk.download('stopwords')\n", "from nltk.corpus import stopwords\n", "from nltk.stem import PorterStemmer" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "S2F_72jaigAq", "outputId": "f990f35d-c979-4f58-b2f7-7b118844d678" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stderr", "text": [ "[nltk_data] Downloading package stopwords to /root/nltk_data...\n", "[nltk_data] Unzipping corpora/stopwords.zip.\n" ] } ] }, { "cell_type": "code", "source": [ "# Cleaning the news\n", "corpus = []\n", "ps = PorterStemmer()\n", "\n", "for i in range(0,news.shape[0]):\n", "\n", " # Cleaning special character from the news-title\n", " title = re.sub(pattern='[^a-zA-Z]', repl=' ', string=news.title[i])\n", "\n", " # Converting the entire news-title to lower case\n", " title = title.lower()\n", "\n", " # Tokenizing the news-title by words\n", " words = title.split()\n", "\n", " # Removing the stopwords\n", " words = [word for word in words if word not in set(stopwords.words('english'))]\n", "\n", " # Stemming the words\n", " words = [ps.stem(word) for word in words]\n", "\n", " # Joining the stemmed words\n", " title = ' '.join(words)\n", "\n", " # Building a corpus of news-title\n", " corpus.append(title)\n" ], "metadata": { "id": "vxJ8PJp4k0Px" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "corpus[0:10]" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "8kqVFk5WmMyX", "outputId": "407d2eca-e04a-414a-b3b4-55082890f3d4" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "['hous dem aid even see comey letter jason chaffetz tweet',\n", " 'flynn hillari clinton big woman campu breitbart',\n", " 'truth might get fire',\n", " 'civilian kill singl us airstrik identifi',\n", " 'iranian woman jail fiction unpublish stori woman stone death adulteri',\n", " 'jacki mason hollywood would love trump bomb north korea lack tran bathroom exclus video breitbart',\n", " 'beno hamon win french socialist parti presidenti nomin new york time',\n", " 'back channel plan ukrain russia courtesi trump associ new york time',\n", " 'obama organ action partner soro link indivis disrupt trump agenda',\n", " 'bbc comedi sketch real housew isi caus outrag']" ] }, "metadata": {}, "execution_count": 16 } ] }, { "cell_type": "code", "source": [ "# Creating the Bag of Words model\n", "from sklearn.feature_extraction.text import CountVectorizer\n", "cv = CountVectorizer(max_features=5000, ngram_range=(1,3))\n", "X = cv.fit_transform(corpus).toarray()" ], "metadata": { "id": "6LlsXB02mSpH" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "X.shape" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "Ps4hJtoGmWdn", "outputId": "0ab845fd-4796-474f-ed28-1bf2cc2192bb" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(18285, 5000)" ] }, "metadata": {}, "execution_count": 18 } ] }, { "cell_type": "code", "source": [ "X[0:10]" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "Ca9qf7OXmaEP", "outputId": "4e5f5e0a-4510-49ed-823e-9b8ac6449a9a" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "array([[0, 0, 0, ..., 0, 0, 0],\n", " [0, 0, 0, ..., 0, 0, 0],\n", " [0, 0, 0, ..., 0, 0, 0],\n", " ...,\n", " [0, 0, 0, ..., 0, 0, 0],\n", " [0, 0, 0, ..., 0, 0, 0],\n", " [0, 0, 0, ..., 0, 0, 0]])" ] }, "metadata": {}, "execution_count": 19 } ] }, { "cell_type": "code", "source": [ "# Extracting dependent variable from the dataset\n", "y = news['label']" ], "metadata": { "id": "EB_Nc5d7md1o" }, "execution_count": null, "outputs": [] }, { "cell_type": "markdown", "source": [ "**Model Building**" ], "metadata": { "id": "E5QYgwLSd7JG" } }, { "cell_type": "code", "source": [ "from sklearn.model_selection import train_test_split\n", "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=0)" ], "metadata": { "id": "xk0wbBQXefDm" }, "execution_count": null, "outputs": [] }, { "cell_type": "markdown", "source": [ "**Multinomial Naive Bayes**" ], "metadata": { "id": "vTcqLRX5ehj1" } }, { "cell_type": "code", "source": [ "# Fitting Naive Bayes to the Training set\n", "from sklearn.naive_bayes import MultinomialNB\n", "nb_classifier = MultinomialNB()\n", "nb_classifier.fit(X_train, y_train)" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 80 }, "id": "rL1cguZlXHTi", "outputId": "9d747603-4a72-4dea-dff7-033c899248ed" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "MultinomialNB()" ], "text/html": [ "
MultinomialNB()
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
" ] }, "metadata": {}, "execution_count": 28 } ] }, { "cell_type": "code", "source": [ "# Predicting the Test set results\n", "nb_y_pred = nb_classifier.predict(X_test)" ], "metadata": { "id": "ZBGsG6GCXMEp" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Accuracy, Precision and Recall\n", "from sklearn.metrics import accuracy_score, precision_score, recall_score\n", "score1 = accuracy_score(y_test, nb_y_pred)\n", "score2 = precision_score(y_test, nb_y_pred)\n", "score3 = recall_score(y_test, nb_y_pred)\n", "print(\"---- Scores ----\")\n", "print(\"Accuracy score is: {}%\".format(round(score1*100,2)))\n", "print(\"Precision score is: {}\".format(round(score2,2)))\n", "print(\"Recall score is: {}\".format(round(score3,2)))" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "6yWHMlm9XM7O", "outputId": "617f3687-270f-4c40-f518-dddd7527ab42" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "---- Scores ----\n", "Accuracy score is: 100.0%\n", "Precision score is: 1.0\n", "Recall score is: 1.0\n" ] } ] }, { "cell_type": "code", "source": [ "# Making the Confusion Matrix\n", "from sklearn.metrics import confusion_matrix\n", "nb_cm = confusion_matrix(y_test, nb_y_pred)\n", "\n" ], "metadata": { "id": "UBwZiIeGXWIt" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "nb_cm" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "kjgGXAhfXb9Z", "outputId": "efc62d98-3375-4814-aca2-8bef773f600e" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "array([[3, 0],\n", " [0, 1]])" ] }, "metadata": {}, "execution_count": 32 } ] }, { "cell_type": "code", "source": [ "# Plotting the confusion matrix\n", "plt.figure(figsize=(10,7))\n", "sns.heatmap(data=nb_cm, annot=True, cmap=\"Blues\", xticklabels=['Real', 'Fake'], yticklabels=['Real', 'Fake'])\n", "plt.xlabel('Predicted values')\n", "plt.ylabel('Actual values')\n", "plt.title('Confusion Matrix for Multinomial Naive Bayes Algorithm')\n", "plt.show()" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 641 }, "id": "cmwb9eXGXmJw", "outputId": "052c0db7-bc07-45bf-c1c8-30415b7857e8" }, "execution_count": null, "outputs": [ { "output_type": "display_data", "data": { "text/plain": [ "
" ], "image/png": "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\n" }, "metadata": {} } ] }, { "cell_type": "code", "source": [ "# Hyperparameter tuning the Naive Bayes Classifier\n", "best_accuracy = 0.0\n", "alpha_val = 0.0\n", "for i in np.arange(0.1,1.1,0.1):\n", " temp_classifier = MultinomialNB(alpha=i)\n", " temp_classifier.fit(X_train, y_train)\n", " temp_y_pred = temp_classifier.predict(X_test)\n", " score = accuracy_score(y_test, temp_y_pred)\n", " print(\"Accuracy score for alpha={} is: {}%\".format(round(i,1), round(score*100,2)))\n", " if score>best_accuracy:\n", " best_accuracy = score\n", " alpha_val = i\n", "print('--------------------------------------------')\n", "print('The best accuracy is {}% with alpha value as {}'.format(round(best_accuracy*100, 2), round(alpha_val,1)))" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "trKpGSt0XvUB", "outputId": "3fe6f325-6779-4495-bf9f-0cb9a6621999" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Accuracy score for alpha=0.1 is: 100.0%\n", "Accuracy score for alpha=0.2 is: 100.0%\n", "Accuracy score for alpha=0.3 is: 100.0%\n", "Accuracy score for alpha=0.4 is: 100.0%\n", "Accuracy score for alpha=0.5 is: 100.0%\n", "Accuracy score for alpha=0.6 is: 100.0%\n", "Accuracy score for alpha=0.7 is: 100.0%\n", "Accuracy score for alpha=0.8 is: 100.0%\n", "Accuracy score for alpha=0.9 is: 100.0%\n", "Accuracy score for alpha=1.0 is: 100.0%\n", "--------------------------------------------\n", "The best accuracy is 100.0% with alpha value as 0.1\n" ] } ] }, { "cell_type": "markdown", "source": [ "**Logistic Regression**" ], "metadata": { "id": "SAM_SPFzX4nV" } }, { "cell_type": "code", "source": [ "from sklearn.linear_model import LogisticRegression\n", "lr_classifier = LogisticRegression(random_state=0)\n", "lr_classifier.fit(X_train, y_train)" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 80 }, "id": "n315pWqQX6Ko", "outputId": "0e7235df-ba37-4aeb-8cc2-3a47c0ac14ad" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "LogisticRegression(random_state=0)" ], "text/html": [ "
LogisticRegression(random_state=0)
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
" ] }, "metadata": {}, "execution_count": 35 } ] }, { "cell_type": "code", "source": [ "# Predicting the Test set results\n", "lr_y_pred = lr_classifier.predict(X_test)" ], "metadata": { "id": "fzy1FjfkX-Rg" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "# Accuracy, Precision and Recall\n", "from sklearn.metrics import accuracy_score, precision_score, recall_score\n", "score1 = accuracy_score(y_test, lr_y_pred)\n", "score2 = precision_score(y_test, lr_y_pred)\n", "score3 = recall_score(y_test, lr_y_pred)\n", "print(\"---- Scores ----\")\n", "print(\"Accuracy score is: {}%\".format(round(score1*100,2)))\n", "print(\"Precision score is: {}\".format(round(score2,2)))\n", "print(\"Recall score is: {}\".format(round(score3,2)))" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "2nifyHQvYCHx", "outputId": "2fbd5b18-47f8-4f1c-b027-f0271d3edb86" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "---- Scores ----\n", "Accuracy score is: 100.0%\n", "Precision score is: 1.0\n", "Recall score is: 1.0\n" ] } ] }, { "cell_type": "code", "source": [ "# Making the Confusion Matrix\n", "from sklearn.metrics import confusion_matrix\n", "lr_cm = confusion_matrix(y_test, lr_y_pred)\n", "\n", "\n", "\n", "lr_cm" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "teHahNZPYIxY", "outputId": "baae6f93-4c26-4234-aa4e-9663df2ac92b" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "array([[3, 0],\n", " [0, 1]])" ] }, "metadata": {}, "execution_count": 38 } ] }, { "cell_type": "code", "source": [ "# Plotting the confusion matrix\n", "plt.figure(figsize=(10,7))\n", "sns.heatmap(data=nb_cm, annot=True, cmap=\"Blues\", xticklabels=['Real', 'Fake'], yticklabels=['Real', 'Fake'])\n", "plt.xlabel('Predicted values')\n", "plt.ylabel('Actual values')\n", "plt.title('Confusion Matrix for Logistic Regression Algorithm')\n", "plt.show()" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 641 }, "id": "DUhK6cwWYNTD", "outputId": "ed82f5f0-be84-4b94-e50b-00c71215891e" }, "execution_count": null, "outputs": [ { "output_type": "display_data", "data": { "text/plain": [ "
" ], "image/png": "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\n" }, "metadata": {} } ] }, { "cell_type": "code", "source": [ "# Hyperparameter tuning the Logistic Regression Classifier\n", "best_accuracy = 0.0\n", "c_val = 0.0\n", "for i in np.arange(0.1,1.1,0.1):\n", " temp_classifier = LogisticRegression(C=i, random_state=0)\n", " temp_classifier.fit(X_train, y_train)\n", " temp_y_pred = temp_classifier.predict(X_test)\n", " score = accuracy_score(y_test, temp_y_pred)\n", " print(\"Accuracy score for C={} is: {}%\".format(round(i,1), round(score*100,2)))\n", " if score>best_accuracy:\n", " best_accuracy = score\n", " c_val = i\n", "print('--------------------------------------------')\n", "print('The best accuracy is {}% with C value as {}'.format(round(best_accuracy*100, 2), round(c_val,1)))" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "39RS_Bo8YUvY", "outputId": "5a47d877-0400-493e-c575-4c7c096eace3" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Accuracy score for C=0.1 is: 75.0%\n", "Accuracy score for C=0.2 is: 75.0%\n", "Accuracy score for C=0.3 is: 75.0%\n", "Accuracy score for C=0.4 is: 100.0%\n", "Accuracy score for C=0.5 is: 100.0%\n", "Accuracy score for C=0.6 is: 100.0%\n", "Accuracy score for C=0.7 is: 100.0%\n", "Accuracy score for C=0.8 is: 100.0%\n", "Accuracy score for C=0.9 is: 100.0%\n", "Accuracy score for C=1.0 is: 100.0%\n", "--------------------------------------------\n", "The best accuracy is 100.0% with C value as 0.4\n" ] } ] }, { "cell_type": "code", "source": [ "classifier = LogisticRegression(C=0.8, random_state=0)\n", "classifier.fit(X_train, y_train)" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 80 }, "id": "s2FwmbJSYZI4", "outputId": "ce7e75a0-0b39-43c4-d600-9f4433b32c3d" }, "execution_count": null, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "LogisticRegression(C=0.8, random_state=0)" ], "text/html": [ "
LogisticRegression(C=0.8, random_state=0)
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
" ] }, "metadata": {}, "execution_count": 41 } ] }, { "cell_type": "markdown", "source": [ "**Predictions**" ], "metadata": { "id": "CzKqHOW6YgRQ" } }, { "cell_type": "code", "source": [ "def fake_news(sample_news):\n", " sample_news = re.sub(pattern='[^a-zA-Z]',repl=' ', string=sample_news)\n", " sample_news = sample_news.lower()\n", " sample_news_words = sample_news.split()\n", " sample_news_words = [word for word in sample_news_words if not word in set(stopwords.words('english'))]\n", " ps = PorterStemmer()\n", " final_news = [ps.stem(word) for word in sample_news_words]\n", " final_news = ' '.join(final_news)\n", "\n", " temp = cv.transform([final_news]).toarray()\n", " return classifier.predict(temp)" ], "metadata": { "id": "kFbdXTmsYhXI" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "from google.colab import drive\n", "drive.mount('/content/drive/')\n", "\n", "df_test = pd.read_csv('/content/kaggle_fake_test.csv')\n", "" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "_5zBI2iHYj8I", "outputId": "bbc5ac85-9758-42f1-c6b4-6973bf0b5ce4" }, "execution_count": 52, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Drive already mounted at /content/drive/; to attempt to forcibly remount, call drive.mount(\"/content/drive/\", force_remount=True).\n" ] } ] }, { "cell_type": "code", "source": [ "df_test.columns" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "udPj4z_rcQM9", "outputId": "e0b08df5-cd84-4516-e92f-a3473ef8459f" }, "execution_count": 53, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "Index(['id', 'title', 'author', 'text'], dtype='object')" ] }, "metadata": {}, "execution_count": 53 } ] }, { "cell_type": "code", "source": [ "news_title = df_test['title']\n", "\n", "\n", "\n", "news_title.shape" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "m0DyPjs4cTVk", "outputId": "ccf5da3d-f9d6-4ba5-ad5e-2e2f62e47e2b" }, "execution_count": 54, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "(5200,)" ] }, "metadata": {}, "execution_count": 54 } ] }, { "cell_type": "code", "source": [ "# For generating random integer\n", "from random import randint\n", "\n", "# Predicting values\n", "row = randint(0,news_title.shape[0]-1)\n", "sample_news = news_title[row]\n", "\n", "print('News: {}'.format(sample_news))\n", "if fake_news(sample_news):\n", " print('Prediction: This is a FAKE news!')\n", "else:\n", " print('Prediction: This is a REAL news.')" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "1Z_t_dctcX60", "outputId": "fcf36c92-ea3e-4569-a7ed-0473805e43b3" }, "execution_count": 55, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "News: Obama Puts Syria at Arm’s Length as Carnage Drags On - The New York Times\n", "Prediction: This is a REAL news.\n" ] } ] }, { "cell_type": "code", "source": [ "# Predicting values\n", "row = randint(0,news_title.shape[0]-1)\n", "sample_news = news_title[row]\n", "\n", "print('News: {}'.format(sample_news))\n", "if fake_news(sample_news):\n", " print('Prediction: This is a FAKE news!')\n", "else:\n", " print('Prediction: This is a REAL news.')" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "brTtKvvQcjck", "outputId": "f2908833-8373-4d4c-dd28-7db808d3a04a" }, "execution_count": 56, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "News: Eight Clowns Jabbering—But Next Time Please Spare Us The Pecksniffery\n", "Prediction: This is a REAL news.\n" ] } ] }, { "cell_type": "code", "source": [ "# Predicting values\n", "row = randint(0,news_title.shape[0]-1)\n", "sample_news = news_title[row]\n", "\n", "print('News: {}'.format(sample_news))\n", "if fake_news(sample_news):\n", " print('Prediction: This is a FAKE news!')\n", "else:\n", " print('Prediction: This is a REAL news.')" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "We4CM3PrcoOc", "outputId": "99372743-1868-4caa-bf98-a6a0f8ec4c80" }, "execution_count": 57, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "News: Ra-Ra-Rasputin: Russia’s famous love machine gets new biography this fall\n", "Prediction: This is a REAL news.\n" ] } ] }, { "cell_type": "code", "source": [ "# Predicting values\n", "row = randint(0,news_title.shape[0]-1)\n", "sample_news = news_title[row]\n", "\n", "print('News: {}'.format(sample_news))\n", "if fake_news(sample_news):\n", " print('Prediction: This is a FAKE news!')\n", "else:\n", " print('Prediction: This is a REAL news.')" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "-zTikfDFcrYd", "outputId": "4fc1e596-0fb9-4948-a61a-b2ae8fb56309" }, "execution_count": 58, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "News: New MH370 analysis suggests no one at controls during crash\n", "Prediction: This is a REAL news.\n" ] } ] } ] }