Ved Prakash Shukla

Conducted sentiment analysis on 34,660 Amazon product reviews from Kaggle.

• Preprocessed data, managed missing values, and corrected class imbalances.
• Developed custom word dictionary with 4,223 features; integrated GloVe embeddings (50-dim).
• Implemented Naive Bayes, SVM, and LSTM models using 60/20/20 train-validation-test split.
• Achieved 71.5% accuracy with LSTM model, surpassing performance of other algorithms.
• Evaluated trade-offs between word count features and embeddings; proposed solutions for imbalance concerns.

Data Analysis & Visualization (Banking Operations)

• Conducted data quality checks for missing values, duplicates, and inconsistencies.
• Applied cleaning techniques and validation rules to ensure data integrity.
• Used statistical and visualization methods for exploratory data analysis.
• Identified customer behavior trends and transaction anomalies as performance indicators.
• Transformed raw data into strategic insights for business decision-making.
• Built dashboards and visual reports in Power BI/Tableau to present findings clearly.
• Delivered presentations to teams, aligning insights with business objectives.