MachineHack recently launched its latest hackathon called Predict The Price Of Books. This article is a complete step by step guide to the solution.
Click here to participate in the hackathon.
Predict The Book Price Hackathon
The so-called paradoxes of an author, to which a reader takes exception, often exist not in the author’s book at all, but rather in the reader’s head. – Friedrich Nietzsche
Books are open doors to an unimagined world which is unique to every person. It is more than just a hobby for many. There are many among us who prefer to spend more time with books than anything else.
Here we explore a big database of books. Books of different genres, from thousands of authors. In this challenge, participants are required to use the dataset to build a Machine Learning model to predict the price of books based on a given set of features.
Size of training set: 6,237 records
Size of test set: 1,560 records
Click here to participate in the hackathon.
FEATURES:
- Title: The title of the book
- Author: The author(s) of the book.
- Edition: The edition of the book eg (Paperback,– Import, 26 Apr 2018)
- Reviews: Customer reviews about the book
- Ratings: The customer ratings of the book
- Synopsis: The synopsis of the book
- Genre: The genre the book belongs to
- BookCategory: The department the book is usually available at.
- Price: The price of the book (Target variable)
Click here to participate in the hackathon.
The following Python notebook contains the complete step by step guide to work on the above-mentioned hackathon. Use this notebook to learn and adapt to this work to better your score.
Approach
- Exploring The Data Sets
- Cleaning, Processing and Generating New Features
- Building A Regressor
- Optimising The Hyperparameters Using Bayesian Optimisation
Getting The Datasets
Go to MachineHack, Sign Up as a user and click on the Predict The Price Of Books Hackathon. Start the hackathon and find the dataset in the Attachment section.
Click here to register for the hackathon
Without further ado, let’s crack the Hackathon!
Click here to participate in the hackathon.
The above solution gives an average score of approximately 65 percent accuracy using RMLSE for evaluation. Use the solution, tweak and tune it to better the score.
Good Luck !
