Search-Engines-Project

In this project, we present a complete pipeline for books recommendation. Given a query from the user, the recommendation system finds the books that best match this query, based on the books abstracts, tags, comments and reviews, and other metrics. This process is divided into several steps. First, we collect more than 20 thousand books manually scraped from GoodReads website to form our dataset. Then, we preprocess the dataset and use Elasticsearch for both storing and processing of the data. Finally, using search and ranking algorithms, we retrieve and score the results, resulting in an engine that produces relatively good results based on the evaluation done. This implementation can therefore be used in a larger extent on bigger datasets to provide the user with book recommendations.