Reinforcement learning-based PoI recommendation using reviews, ratings and locations data

The growth in internet data creates a lot of issues in decision-making for tourists to find their favorite places. To improve travel services, many internet platforms started gathering feedback from visitors in the form of numeric ratings and reviews. The numeric rating may not be a clear parameter to find the actual rating, and in such cases, a review may help in better understanding the real feedback for future travelers. The combined numeric and review data can be used to develop a better recommendation system to provide the top tourist spots based on the feedback. This research paper aims to build on a selected premise to develop a recommendation system that can suggest and map out an entire journey around a city curated in a personalized way. In order to find recommended tourist places, we use a reinforcement learning (RL) algorithm to understand the specific needs of a tourist and create a map of destinations around a selected city. The proposed model provides the suggested locations with better accuracy based on travelers’ needs.

This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0).