Creation of semantic location profiles using Bayes, rule-based, trees and meta classification approaches

Travel recommender systems are gaining a higher popularity in the society due to their capability of planning trips in a short time period. The challenge of providing the most accurate recommendations has become a complex and difficult task due to the numerous variations in user preferences. In order to provide the most accurate recommendations, it is necessary to consider additional parameters like the prevailing weather condition, which has a direct influence on the user preference to a particular location. Therefore, when providing travel recommendations considering the weather context, it was identified that the ability to correctly identify a location as an indoor or outdoor attraction plays a vital role in improving the accuracy of the recommendations. Considering the millions of locations available in Google Places, it is difficult to manually tag the location status as indoor or outdoor. This paper provides a novel approach to determine the status of a particular location based on the identified attributes of the location. Moreover, the experimental results and the accuracy of the predicted outputs have been discussed by using different classifiers under the Bayes, Rule-Based, Trees and Meta classification approaches.