README.md

---
task_categories:
- fill-mask
language:
- en
tags:
- agriculture Extension
pretty_name: AEC1.1
configs:
- config_name: default
  data_files:
  - split: train
    path: "ae_corpus_v1_train_partial.txt"
  - split: validation
    path: "ae_corpus_v1_dev_partial.txt"
  - split: test
    path: "ae_corpus_v1_test_partial.txt"
---

# DISCLAIMER: DUE TO AN ONGOING PUBLICATION REVIEW FOR ITS ASSOCIATED JOURNAL PAPER, THIS REPO ONLY CONTAINS PARTIAL DATA SPLITS. THE REST OF THE DATA WILL BE UPLOADED UPON ACCEPTANCE. 

# Dataset Card for AEC


## Dataset Description

The Agricultural Extension Corpus 1.1 is a compilation of 1655 official agricultural Extension documents (e.g., fact sheets, digital books) concerning water-related and sustainable agricultural practices research. 

- **Homepage:** https://huggingface.co/datasets/msu-ceco/aec_v1
- **Paper:** [More Information Needed]
- **Curated by:** [DSI Lab](https://dsiweb.cse.msu.edu/)
- **Point of Contact:** Dr. A.Pouyan Nejadhashemi (pouyan@msu.edu)
- **Language(s) (NLP):** English (`en`)
- **License:** [More Information Needed]


## Uses

<!-- Address questions around how the dataset is intended to be used. -->

### Direct Use

It is intended to train ML models in general and for NLP tasks in particular (e.g., Masked Language Modeling).


### Out-of-Scope Use

This dataset should not be used as a reference/answer to address or respond to time-sensitive and location-sensitive questions.


## Dataset Structure

<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->
This dataset does not contain any specific fields. It is a corpus of paragraphs and sentences that has been split into training, validation, and test sets using an 80-18-2 ratio


## Dataset Creation

### Curation Rationale

The motivation for the creation of this dataset is to provide practitioners of applied AI in agricultural fields with a reliable base corpus that can be used for several NLP downstream tasks (MLM, NER, QA, etc.).


### Source Data

Two categories of sources:
 - [The Journal Of Extension](https://archives.joe.org/)
 - Land-grant institutions’ official Extension websites (e.g., [MSU Extension](https://www.canr.msu.edu/outreach/index), [UNL Extension](https://extension.unl.edu/), etc.)
   - We covered at least one Extension website in each of the 40 (out of 48) states in the continental USA. Due to the limitations of appropriate software, we only provided author references categorized per all documents found in each state. The list of states and the references can be found [here](https://huggingface.co/datasets/msu-ceco/aec_v1/blob/main/All_References_APA_AEC1.1.docx).

#### Data Collection and Processing

<!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->
The dataset was compiled from publicly available documents: (1) 212 scholarly full-text articles from the [Journal Of Extension](https://archives.joe.org/) and (2) 1443 materials extracted from 40 land-grant institutions that provide Extension services throughout the USA. 
These materials include factsheets, e-books, and articles related to agricultural research, and they are in HTML, text, and PDF formats.

The PDFs were processed and converted into text using [Amazon's Textract](https://aws.amazon.com/textract/ocr/).

Furthermore, we removed non-UTF8 characters, bibliographic references, and URLs (when possible). 


#### Who are the source data producers?

<!-- This section describes the people or systems who originally created the data. It should also include self-reported demographic or identity information for the source data creators, if this information is available. -->
The original documents were authored by Extension staff (educators, faculty, specialists, etc.) from diverse land-grant institutions. The references to these authors can be found [here (TTO-DO)]().


#### Personal and Sensitive Information

<!-- State whether the dataset contains data that might be considered personal, sensitive, or private (e.g., data that reveals addresses, uniquely identifiable names or aliases, racial or ethnic origins, sexual orientations, religious beliefs, political opinions, financial or health data, etc.). If efforts were made to anonymize the data, describe the anonymization process. -->
To the best of our ability, we tried removing any personal information (e.g., emails) during the automatic processing stage. However, considering that this dataset was programmatically compiled from other existing texts, there could be residues of personal information, (especially from bibliographical sections) that could have carried over into our dataset as well. 
If you are the original author of a document compiled inside AEC and found your personal details in the dataset, please refer to our list of references [here TO_DO](). If you have not been cited, please contact us to request removal.


## Bias, Risks, and Limitations

<!-- This section is meant to convey both technical and sociotechnical limitations. -->

[More Information Needed]

### Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

Users should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.

## Citation

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**

[More Information Needed]

**APA:**

[More Information Needed]