文件列表:

Code (0, 2021-01-05)
Code\Polarization_Graphs.ipynb (33289, 2021-01-05)
Code\Political_Bias_Classifier.ipynb (36078, 2021-01-05)
Data (0, 2021-01-05)
Data\data_locations.md (185, 2021-01-05)

# Quantifying-Political-Polarization Dataset abtracted from https://www.kaggle.com/snapcrack/all-the-news This project was my undergraduate thesis that had two main goals - build an accurate classifier of news article political bias and use said classifier to see how polarization occurred in major news sources over the course of the 2016 election period. ### Project TL;DR: Background - Political polarization in the US has been becoming an increasingly large threat to the American democracy, particularly during recent election periods. Giving users a way of estimating the bias of their news article allows them to "take it with a grain of salt". Method - First, pre process the articles. Then, feed the processed articles into a pipeline that vectorizes and transforms the text into numerical values. Next, use a Support Vector Machine (SVM) to classify the articles' political bias accurately. Finally, use said trained classifier to analyze how major news sources' political bias changed from 2016 to 2017. Results - The SVM algorithm was able to achieve ~ 80% accuracy between a 5 label selection. Some news sources had statistically significant changes in their political biases over the course of the 2016 election period. ### Folder Contents: Code: Political_Bias_Classifier - Processes and classifies the articles into 5 categories Polarization_Graphs - Uses the classifier to identify trends in major news sources over time Data: Links to the data for Political_Bias_Classifier (Classifier_Data.csv) and Polarization_Graphs (Trend_Data.csv)

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：