In this project we take the input of twitter data in three languages English, German and Russian. This data is being indexed in Solr and we then apply three different IR Machine Learning models :
- Language Model
- BM25
- Divergence and Randomness Model(DFR)
Then for each model we evaluate the performance using TREC_eval for the given queries. In later stage we use MAP values to compare the performance of all the three IR models implemented.