AI Generated Text Classfier and Text Similarity Finding

Data collection

Wiki data collection

Reverse data collection

US Election 2024

Get news links from different sources

Extract data from the links

Find the topics (get words)

Ask ChatGPT to form questions using those words

Collect the ChatGPT answers and use as AI generated text

Calculate text similarity

Features with TF-IDF

Features without TF-IDF

Features with TF-IDF + PCA (with 1024 features)

Cosine similarity

Jaccard similarity

Compare

Data preprocessing

Feature engineering

LDA Topic modeling

Save all the features

Exploratory Data Analysis

Machine Learning model

Random forest

Support Vector Machine

XGBoost

Deep Learning model

BiLSTM

Hybrid model

Explanation

Using SK Learn feature importance

SHAPE for feature explaination