Travel Direction Recommendation Model Based on Photos of User Social Network Profile

Abstract

Travelling is one of the most enjoyable activities for people of all ages. It is constantly looking for innovative solutions on how to tailor travel recommendations to the needs of its customers. The purpose of our proposed recommendation model is to suggest travelling countries based on photos from the user's social network account and metadata associated with the photos. Such recommendation models are highly dependent on the data used in the model preparation steps and on the technologies and methods implemented in the model. The newly collected data from the Instagram users' accounts were used in the model preparation. The recommendation system is based on the combination of four methods: object detection, similarity measures, classification, and data clustering. The novelty of the proposed recommendation model is that it adopts different data (Instagram photos) for travel direction recommendation, defines a new combined method, integrates results of similarity measurement and SOM application results into one final recommendation, and estimates the parameter impact for different components of recommendation model. A proposed evaluation measure has been used to conclude the results of the recommendation model and as a result the names of the travelling countries have been recommended. The results of the proposed recommendation model are promising, and the validation results demonstrate that on average 63% of the users who visited countries match the recommendations provided for the trip directions, while the accuracy of recommendations, matching user visited countries, but not presented in the photos for recommendation estimation, on average was 96%. The accuracy performance is very positive, while the recommendation system is fully automated and machine learning based. With time, the accuracy of the model may even increase by adopting the photo metadata (location).INDEX TERMS classification, object detection, recommendation model, self-organizing maps, similarity measure, photo of social networks.

The Efficiency of YOLOv5 Models in the Detection of Similar Construction Details

“…For example, in medicine, image data can be used to predict different diseases, such as cancer [1,2], glaucoma [3,4], and pneumonia [5,6]. Object detection models can be used in systems for travel direction recommendation [7], in industry for solutions to robotization tasks [8,9], in face detection for different applications [10,11], or other fields [12][13][14][15][16]. Usually, in all research, various computer vision methods or combinations are used to solve the specific problem.…”