Mukul Bichkar

The idea of the project is to build information retrieval systems, evaluate and compare their performance levels in terms of retrieval effectiveness. The dataset provided to us contains 3204 raw documents. We used CACM test collection as our textual corpus, 64 unprocessed queries, Relevance judgments and a stop-list.

• Developed search engines with BM25, Lucene, tf.idf, query likelihood as retrieval models.
• Performed Query Expansion techniques on BM25 and tf.idf using Pseudo-Relevance Feedback approach.
• Performed Stopping, Stemming and indexed the stemmed version of the corpus that was provided to us and performed query by query analysis on all the models.
• A total of 11 distinct runs were generated to assess the performance of these search engines.

1. Mean Average Precision
2. Mean Reciprocal Rank
3. p@K, K = 5 and 20
4. Precision and Recall

Designed a mobile first interactive website that allows users to search for the favorite recipes by name, cuisine and by the ingredients available in their kitchen. The app provides detailed recipe procedure for selected recipe with speech and video assistance.

It's a MEAN Stack application which is mobile responsive.

In this application one can search for a new car, obtain information related to pricing along with year wise analysis, read reviews about cars from experts as well as other users, get information about dealers, repairshops and their reviews. Users' can register themselves, login, follow other users, write reviews about cars.