For skill demonstration this page contains both github link as well as actuall project links
Sentiment Analysis
Built an application of sentiment classification base on Twitter data with the help of transfer learning models from hugging face, also used ensemble technique to get the best results. Used the following model in the ensemble ['ProsusAI/finbert', 'finiteautomata/bertweet-base-sentiment-analysis', 'elozano/tweet_sentiment_eval']
Attrition Rate
Performed some data analysis on a Kaggle based dataset. Applied some features transformation, outlier removal at last made a ML model pipeline base on VotingClassifier ( accuracy 89%, F1 score 0.93)
Project GitHub link
Some Machine Learning models from scratch
Built few ML models from scratch using NumPy, TensorFlow
- Linear Regression
- Logistic Regression
- Anamoly Detection
- K-Means Clustering
- Neural Network
- Collaborative Filtering Recommender System