ABSTRACT
Rapid developments in the usage of Internet enabled the need of predictive analytics and assistance to understand the behavior of buyers and sellers on e-commerce platform. Online shopping has evolved in many years since its inception, and also inherited lot of advancements in the way we live, shop, and do business. Finding potential new buyers, retargeting the existing one, handling queries, and maintaining the inventories are very challenging modules and will affect the e-commerce industry very badly if not taken care of. Artificial Intelligence (AI) is being used by top e-commerce industries like Amazon and Flipkart to understand the behavior of the buyer by introducing chatbots and customized recommendations by predicting the shopping pattern of the buyers. These top industries maintain and process millions of buyer behaviors to predict and target the right customers. A novel technique called Call based Intelligent Bot Personal Assistance (CIB-PA) using Machine Learning (ML) and Natural Language Processing (NLP) 62has been proposed in the paper to self-serve the seller in assisting and targeting the right buyer more accurately. CIB-PA is deployed by integrating text analysis via call recording, Intelligence as a Service (IasS), Personal Assistance, and Bot Integrations. Test bed was deployed using App development toolkits, Hadoop MapReduce, DataRobot, RapidMiner, Fusioo, BigML, and CoreNLP. Through simulations, it is observed that CIB-PA outperforms in predicting buyer behavior.
