Fake News Detection Using BERT
Developed an advanced Natural Language Processing (NLP) system to detect fake news articles using BERT (Bidirectional Encoder Representations from Transformers). The objective was to classify news content as genuine or misleading with high accuracy.
Collected and preprocessed textual data by cleaning, tokenizing, and preparing it for transformer-based modeling. Fine-tuned a pre-trained BERT model on a labeled news dataset to improve classification performance.
Evaluated the model using accuracy, precision, recall, and F1-score, achieving robust results in distinguishing fake and authentic news. This project strengthened my expertise in deep learning, NLP, and transformer architectures.
Customer Churn Prediction Using Machine Learning
Developed a machine learning model to predict customer churn using historical customer data. The project aimed to identify customers who were likely to discontinue a service, enabling businesses to take proactive retention measures.
Performed comprehensive data preprocessing, including handling missing values, encoding categorical variables, feature scaling, and exploratory data analysis. Addressed class imbalance using SMOTE to improve model performance on minority classes.
Built and evaluated multiple classification models, including Logistic Regression and Decision Tree, using metrics such as accuracy, precision, recall, and F1-score. The final model successfully identified high-risk customers and provided actionable business insights.
Alibaba RFQ Scraper
Built an ETL-style scraping pipeline using Python, Selenium, and BeautifulSoup to extract RFQ data for market and price analysis; automated multi-page crawling and cleaned/exported structured datasets to CSV; performed EDA to identify vendor patterns and product trends.
Customer Churn Prediction
Built a churn prediction pipeline using Python and Scikit-learn with preprocessing, encoding, and SMOTE balancing; trained a Decision Tree model with 77% accuracy and AUC 0.73; deployed a real-time Streamlit app and implemented feature engineering for improved model interpretability and churn insights.
