Abstract
Many personality theories suggest that personality influences customer shopping preference. Thus, this research analyses the potential ability to improve the accuracy of the collaborative filtering recommender system by incorporating the Five-Factor Model personality traits data obtained from customer text reviews. The study uses a large Amazon dataset with customer reviews and information about verified customer product purchases. However, evaluation results show that the model leveraging big data by using the whole Amazon dataset provides better recommendations than the recommender systems trained in the contexts of the customer personality traits.
Highlights
Recommender systems (RSs) are nowadays a very important element that is influencing customer digital experience in electronic services
The main goal of the experiment conducted in this study was to integrate the information contained in the users’ text reviews into a RS, and in particular, investigate whether Five-Factor Model (FFM) personality traits, as reflected in the text generated by users, would allow improving the Root Mean Square Error (RMSE) of predicted ratings
The second approach aimed to investigate whether FFM personality traits, as reflected in the text generated by users, would allow improving the Root Mean Square Error (RMSE) of predicted ratings
Summary
Recommender systems (RSs) are nowadays a very important element that is influencing customer digital experience in electronic services. Many major companies such as Amazon, Netflix, or Spotify are successfully employing effective RSs in their businesses and are seeking to improve their algorithms even further. Personality theories researchers claim that human personality traits have a significant influence on customer preferences and subsequently on behavior [7], [8]. They seem to be a promising predictor of customer behavior. It is especially important in digital markets where customer personality characteristics can be inferred from their digital footprints [9], [10]
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have