Predictive and Sentiment Analysis on Factors Influencing E-Commerce Growth in Transition Economies

Abstract

With emerging technology in Data Science, every industry is seen to adapt technological trends to benefit their businesses. It’s a competition on delivering better quality and high value goods and services to the market and this is only possible through deep analysis in understanding consumer demands and market trends. This research aimed at analyzing the e-commerce industry on how to increase growth through identifying factors that influence online customers purchase intentions and used machine-learning algorithm in predicting online sales. The variable “number of orders” has been used as the dependent variable, which is used in determining the growth of e-commerce from the increase in online sales. The research identified factors such as online reviews, price and product quality (ratings) as main factors pushing online purchases in exploring their impact towards the dependent variable. This research involves three main types of analysis; diagnostic analysis, sentiment analysis and predictive analysis. Findings revealed that online consumers are more drawn to making online purchases if the product description, price, reviews and delivery of products and services match with their needs. The sentiment analysis was able to connect the findings from descriptive analysis by understanding consumers thoughts and discussions on the products through the online reviews. Additionally ANN model was able to accurately predict how number of orders is influenced by factors such as price, online reviews, wishlist, product description, rating, number of feedback, delivery time, delivery price, payment methods, stock available and discount price.