Fake News Risk Identifier
End Goal: Train a model to scan a website’s text and predict (0-1) if the website contains fake news, misinformation, or disinformation.
Use: A tool to help judge how much a reader should trust a website.
Problem solved: People blindly trusting information found on websites that may be misleading or give false claims.
Plan: Python TensorFlow through GitHub Build a supervised ML ensemble logistic regression to produce a risk number for false information in the news website being read. This is a natural language processing problem so the documents will be cleaned up into one solid page then data points (#of verbs, punctuation, author’s other written pages etc.) will be extracted. The data will be split 70/30 and the trained model accuracy rate (takes both the false positive and the false negative observations into account) will be aimed for 85% or greater.
- Data collection a. https://drive.google.com/file/d/1IoTRrJNDJqvaG3hnUpnHQyGvPAJbO8y3/view b. https://www.kaggle.com/datasets/jruvika/fake-news-detection c. https://www.hindawi.com/journals/complexity/2020/8885861/#materials-and-methodshttps://www.hindawi.com/journals/complexity/2020/8885861/#materials-and-methods
- Data Cleaning and formatting a. Reduce all pages to one page b. Bring values like title, author, etc. out c. Remove formatting d. Remove pages with <20 words
- Extract Lingustic variables a. Reasurch important variables
- Choose features/variables to train on
- Split training and testing data Repeat until 85% accuracy is achieved - Train model using training data Evaluate model Hyperperameters and model tuning User website >> trained model >> Output (Classification * certainty)
The Model Boosting Classifier (XGBoost) ensamble method Detailed paper on XGBoost - https://dl.acm.org/doi/epdf/10.1145/2939672.2939785 ABSTRACT Tree boosting is a highly effective and widely used machine learning method. In this paper, we describe a scalable end-to-end tree boosting system called XGBoost, which is used widely by data scientists to achieve state-of-the-art results on many machine learning challenges. We propose a novel sparsity-aware algorithm for sparse data and weighted quantile sketch for approximate tree learning. More importantly, we provide insights on cache access patterns, data compression and starting to build a scalable tree boosting system. By combining these insights, XGBoost scales beyond billions of examples using far fewer resources than existing systems.
https://docs.google.com/presentation/d/1Z2e6eevhT8K1k-RBiA-7mnOeBslRVWdWbyYz90iVpec/edit?usp=sharing