AI Uses Goodreads to Predict Amazon Best Sellers

first_imgLet us know what you like about Geek by taking our survey. There are no shortage of instructions for propelling your book to the top of the Amazon charts.Use keywords, create an attractive cover, choose the right category. Have it professionally edited and properly formatted. Write a good description, price it right, get positive reviews.Or, stick your story into an algorithm that can predict whether a book will become a best seller on Amazon within 15 days of its publication.AdChoices广告Researchers at Northwestern University, Microsoft Research India, and the Indian Institute of Technology Kharagpur developed a model that does just that: by studying reading behavior on online database Goodreads.“We have been working on analyzing the popularity dynamics of various social media entities, such as hashtags in Twitter, topics in Quora, etc.,” Animesh Mukherjee, associate professor of computer science at IIT Kharagpur, told TechXplore. “We felt that a similar approach could be taken to analyze the popularity of books and we found Goodreads to be ideal for this investigation.”The social cataloging website—acquired by Amazon in 2013—allows users to search for books, annotations, and reviews and generate library catalogs and reading lists, among other functions.Much to authors’ chagrin, there is no magic spell you can cast over a book to ensure chart-topping success. Popularity can be measured using several parameters—but in this case, researchers focused on how book reading characteristics influence its favor.The team analyzed the collective reading behavior of Goodreads consumers, and studied characteristics of Amazon best sellers in terms of the site’s entities—books, authors, shelves, genres, user status posts.They then developed a machine learning-based model using these unique features to predict whether a book will hit No. 1 in two weeks of its launch.“We observed that the ratings and reviews received by a book on Goodreads are not as effective in predicting the best sellers as the reading status post patterns of the users,” according to Mukherjee.One of my favorite aspects of the site is the ability to post updates about what page I’m on and how much of a book I’ve read (I’m a slow reader, so sue me). Folks can also post comments about their experience with the novel, memoir, cookbook, comic, whatever.“We find these features to be very effective in predicting whether the book is going to be a bestseller in the future,” Mukherjee said.The AI boasts an average accuracy of 88.72 percent in predicting imminent chart-toppers. That’s a 16.4 percent improvement over baseline methods that use traditional popularity factors like ratings or reviews.“One of the most important insights that we obtain from this study is that … a large majority of [Amazon bestsellers] have reader status post patterns that strongly distinguish them from the rest of the books,” Mukherjee said.The study is available to read online via the Cornell University Library.Artificial intelligence is getting smarter every day: It recently detected more mysterious cosmic radio bursts, learned to create drugs from scratch, and helped map developing brains. Read more about algorithms and books on Geek.com. McDonald’s Plans to Serve AI Voice Technology at Drive ThruCIMON Returns to Earth After 14 Months on ISS center_img Stay on targetlast_img read more

Recent Comments