Predicting Purchase Intention Using Twitter Data

Research Square

Abstract

Abstract Recently, there has been a significant rise in the ecommerce industry with the pandemic entailing lockdowns. More and more people have started posting online about the products they want to buy or asking whether they should buy the product or not. There has been a great deal of research going on in trying to figure out the buying patterns of a user and more importantly the factors which determine whether the user will buy the product or not. One such platform is Twitter which has become quite popular in recent years. This study explores the problem of identifying and predicting the purchase intention of a user for a product. Employment of various text analysis models to tweets data reveals we can predict if a user has shown purchase intention towards a product or not. While there are lexicon and Machine learning based approaches to handle this task, our deep learning approach is a new take on purchase intention prediction from social media data. An accuracy of 85.7% was achieved by our model, applying the LSTM Neural Net with data in the Global Vectors (GloVE) embeddings format.

Authors 1

1. Syed Zain Zaidi
   Affiliation as printed

   Delta Galil Industries