|
"""Spambase: A Census Dataset""" |
|
|
|
from typing import List |
|
|
|
import datasets |
|
|
|
import pandas |
|
|
|
|
|
VERSION = datasets.Version("1.0.0") |
|
|
|
DESCRIPTION = "Spambase dataset from the UCI ML repository." |
|
_HOMEPAGE = "https://archive.ics.uci.edu/ml/datasets/Spambase" |
|
_URLS = ("https://archive.ics.uci.edu/ml/datasets/Spambase") |
|
_CITATION = """ |
|
@misc{misc_spambase_94, |
|
author = {Hopkins,Mark, Reeber,Erik, Forman,George & Suermondt,Jaap}, |
|
title = {{Spambase}}, |
|
year = {1999}, |
|
howpublished = {UCI Machine Learning Repository}, |
|
note = {{DOI}: \\url{10.24432/C53G6X}} |
|
}""" |
|
|
|
|
|
urls_per_split = { |
|
"train": "https://huggingface.co/datasets/mstz/spambase/raw/main/spambase.data" |
|
} |
|
features_types_per_config = { |
|
"spambase": { |
|
"word_freq_make": datasets.Value("float64"), |
|
"word_freq_address": datasets.Value("float64"), |
|
"word_freq_all": datasets.Value("float64"), |
|
"word_freq_3d": datasets.Value("float64"), |
|
"word_freq_our": datasets.Value("float64"), |
|
"word_freq_over": datasets.Value("float64"), |
|
"word_freq_remove": datasets.Value("float64"), |
|
"word_freq_internet": datasets.Value("float64"), |
|
"word_freq_order": datasets.Value("float64"), |
|
"word_freq_mail": datasets.Value("float64"), |
|
"word_freq_receive": datasets.Value("float64"), |
|
"word_freq_will": datasets.Value("float64"), |
|
"word_freq_people": datasets.Value("float64"), |
|
"word_freq_report": datasets.Value("float64"), |
|
"word_freq_addresses": datasets.Value("float64"), |
|
"word_freq_free": datasets.Value("float64"), |
|
"word_freq_business": datasets.Value("float64"), |
|
"word_freq_email": datasets.Value("float64"), |
|
"word_freq_you": datasets.Value("float64"), |
|
"word_freq_credit": datasets.Value("float64"), |
|
"word_freq_your": datasets.Value("float64"), |
|
"word_freq_font": datasets.Value("float64"), |
|
"word_freq_000": datasets.Value("float64"), |
|
"word_freq_money": datasets.Value("float64"), |
|
"word_freq_hp": datasets.Value("float64"), |
|
"word_freq_hpl": datasets.Value("float64"), |
|
"word_freq_george": datasets.Value("float64"), |
|
"word_freq_650": datasets.Value("float64"), |
|
"word_freq_lab": datasets.Value("float64"), |
|
"word_freq_labs": datasets.Value("float64"), |
|
"word_freq_telnet": datasets.Value("float64"), |
|
"word_freq_857": datasets.Value("float64"), |
|
"word_freq_data": datasets.Value("float64"), |
|
"word_freq_415": datasets.Value("float64"), |
|
"word_freq_85": datasets.Value("float64"), |
|
"word_freq_technology": datasets.Value("float64"), |
|
"word_freq_1999": datasets.Value("float64"), |
|
"word_freq_parts": datasets.Value("float64"), |
|
"word_freq_pm": datasets.Value("float64"), |
|
"word_freq_direct": datasets.Value("float64"), |
|
"word_freq_cs": datasets.Value("float64"), |
|
"word_freq_meeting": datasets.Value("float64"), |
|
"word_freq_original": datasets.Value("float64"), |
|
"word_freq_project": datasets.Value("float64"), |
|
"word_freq_re": datasets.Value("float64"), |
|
"word_freq_edu": datasets.Value("float64"), |
|
"word_freq_table": datasets.Value("float64"), |
|
"word_freq_conference": datasets.Value("float64"), |
|
"char_freq_;": datasets.Value("float64"), |
|
"char_freq_(": datasets.Value("float64"), |
|
"char_freq_[": datasets.Value("float64"), |
|
"char_freq_!": datasets.Value("float64"), |
|
"char_freq_$": datasets.Value("float64"), |
|
"char_freq_#": datasets.Value("float64"), |
|
"capital_run_length_average": datasets.Value("float64"), |
|
"capital_run_length_longest": datasets.Value("float64"), |
|
"capital_run_length_total": datasets.Value("float64"), |
|
"is_spam": datasets.ClassLabel(num_classes=2, names=("no", "yes")) |
|
}, |
|
|
|
} |
|
features_per_config = {k: datasets.Features(features_types_per_config[k]) for k in features_types_per_config} |
|
|
|
|
|
class SpambaseConfig(datasets.BuilderConfig): |
|
def __init__(self, **kwargs): |
|
super(SpambaseConfig, self).__init__(version=VERSION, **kwargs) |
|
self.features = features_per_config[kwargs["name"]] |
|
|
|
|
|
class Spambase(datasets.GeneratorBasedBuilder): |
|
|
|
DEFAULT_CONFIG = "spambase" |
|
BUILDER_CONFIGS = [ |
|
SpambaseConfig(name="spambase", |
|
description="Spambase for binary classification.") |
|
] |
|
|
|
def _info(self): |
|
info = datasets.DatasetInfo(description=DESCRIPTION, citation=_CITATION, homepage=_HOMEPAGE, |
|
features=features_per_config[self.config.name]) |
|
|
|
return info |
|
|
|
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]: |
|
downloads = dl_manager.download_and_extract(urls_per_split) |
|
|
|
return [ |
|
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloads["train"]}) |
|
] |
|
|
|
def _generate_examples(self, filepath: str): |
|
data = pandas.read_csv(filepath) |
|
|
|
for row_id, row in data.iterrows(): |
|
data_row = dict(row) |
|
|
|
yield row_id, data_row |
|
|
