Location Aspect Based Sentiment Analyzer for Hotel Recommender System

Abstract

Recently personal recommender system has spread fast, because of its role inhelping users to make their decision. Location-based recommender systems are oneof these systems. These systems are working by sensing the location of the personand suggest the best services to him in his area. Unfortunately, these systems thatdepend on explicit user rating suffering from cold start and sparsity problems. Theproposed system depends on the current user position to recommend a hotel to him,and on reviews analysis. The hybrid sentiment analyzer consists of supervisedsentiment analyzer and the second stage is lexicon sentiment analyzer. This systemhas a contribute over the sentiment analyzer by extracting the aspects that usershave been mentioned in their reviews like (cleanness, service, etc.) by usingaccurate parsing system built on latent semantic analysis results. The accuracymeasurements of the proposed sentiment analyzer were perfect