Internship tasks completed as part of the Machine Learning Internship Program at Cognifyz Technologies (April – May 2025).
All tasks use a restaurant dataset (9,551 records) and are implemented in Python using Scikit-learn and Pandas.
Built a Linear Regression model to predict the aggregate rating of a restaurant based on features such as cuisine type, city, price range, and number of votes.
Preprocessed data via missing value handling, label encoding, and train-test splitting.
📊 Achieved R² score of 0.97 on the test set. Top influencing features identified via coefficient analysis.
Developed a content-based filtering recommendation engine that suggests restaurants based on user preferences — cuisine type, city, and price range.
Includes an interactive widget interface for inputting preferences and visualising filtered results dynamically.
Trained a Random Forest classifier (One-vs-One strategy) combined with TF-IDF text vectorisation (5,000 features) to classify restaurants into cuisine categories.
Evaluated using accuracy, precision, and recall metrics.
🎯 Achieved 92.14% classification accuracy on held-out test data.
| Tool | Usage |
|---|---|
Python |
Core language |
Pandas |
Data loading, cleaning, preprocessing |
Scikit-learn |
ML models, TF-IDF, evaluation metrics |
Matplotlib / Seaborn |
Data visualisation |
ipywidgets |
Interactive recommendation UI (Task 2) |
All tasks use the Zomato Restaurant Dataset — a publicly available dataset containing restaurant details such as name, location, cuisine, pricing, ratings, and votes across multiple cities and countries.
Internship completion certificate available on LinkedIn