More

    BookBee – Book Recommendation System for the welfare of learners

    Introduction 

    BookBee mainly deals with the design of the Book Recommendation System for the welfare of the learners. The current recommendation system lacks many aspects which are found in our research and most of these systems are established with a business point of view. There are many book clubs but there are very few with respect to technical and academia studies. BookBee aims to create a single platform where recommendation of books to students of every stream will be provided with the help of machine learning algorithms such as KNN Algorithm (i.e. k-nearest neighbors). From this portal a Book-Bee community will be organized which will further aid the rating and recommendations of books.

    If you don’t like to read, you haven’t found the right book.

    J.K. Rowling

    Drawbacks Of Existing System

    We have gone through brief amount of research papers about book recommendations and we Observed that the majority uses Content Based Filtering (CBF) which lacks in suggesting proficient books on the basis of user ratings and user reviews and thus helps minimum in decision making process. CBF requires more computing power . Each item must be analyzed for its features, user models must be built, and similarity calculations must be performed. If there are many users and many items, these calculations require significant resources. The weakness of CBF is its low serendipity and overspecialization leading it to recommend items as similar as possible to the ones a user already knows .

    CBF also ignores quality and popularity of items . For instance, two research papers may be considered equally relevant by a CBF recommender system if the papers share the same terms with the user model. This relevance might not always be justified, for example if one paper was written by an authority in the field and presents original results, while another paper was written by a student who paraphrases the results of other research papers. Ideally, a recommender system should recommend only the first paper but a CBF system would fail to do so. In simple words we can draw the conclusion that an alternate approach has to be found which is the main crux point of our project.

    Proposed System

    Current book recommendation systems lack in suggesting proficient books on the basis of user ratings and user reviews and thus helps minimum in the decision-making process. It is using Collaborative Filtering (CF) Algorithm. The main idea of this system is that users like what Like-minded users like, where two users were considered likeminded when they rated items alike. When like-minded users were identified, items that one user rated positively were recommended to the other user, and vice versa. Compared to CBF, CF offers three advantages. First, CF is content independent, If an item is new in the system and has not been rated yet by at least one user, it cannot be recommended.

    In a new community, no users have rated items, so that no recommendations can be made and as a result, the incentive for users to rate items is low. To overcome the cold-start problem, implicit ratings may be inferred from the interactions between users and items. Implicit ratings from the number of pages the users read: the more pages users read, the more the users were assumed to like the books.

    Explanation

    So What our system does is whenever an user selects his subject our model takes an imaginary book of maximum popularity and 5 star rating into consideration. Then using K.N.N. algorithm it searches for ‘K’ nearest neighbours (books) which are approximately similar to it and then after calculating the result top 3 books out of the ‘K’ nearest neighbors are being displayed in the front end of our website.

    This similarity is calculated using cosine similarity approach. In this approach each book is imagined as a vector

    Example: Consider 2 books ‘Learning Python’ and ‘ Python Cookbook’ as 2 vectors and they have θ (angle) of 20 between them then the Similarity score will be cos(20) i.e 93% (approximately similar).

    Conclusion

    It was seen that there are many gaps in Book Recommendation systems. Our solution BookBee seems to fulfil those gaps. It is expected that through BookBee an effective platform would be provided to learners and in particular to academics. BookBee hopes to save precious time for learners thereby improving their efficiency. It would also be beneficial for authors and writers and will help them to emerge.

    With this BookBee hopes to establish a large community of readers, learners, creators and experts. Proper utilization of large data is extremely essential and should be carried out in an efficient manner. The procedures carried out during the project helps in minimizing complications and ambiguity by reducing it to a visual and comparative study.

    Written By:

    • Saurabh Dubey
    • Vivek Gupta
    • Eshaan Kushwaha
    3.8 4 votes
    Article Rating
    Subscribe
    Notify of
    guest
    0 Comments
    Inline Feedbacks
    View all comments

    Recent Articles

    Stay on top - Ge the daily news in your inbox

    Related Stories