AI Supported Smart Service Recommendation Algorithm

Öz

Armut Technology is an online platform that brings together customers and service providers, and positions service providers as business partners with the principle of "Crowdsourcing". Nearly 4000 services are offered within the company. This number is increasing gradually as new service requests are also received from service providers. When customers login to the website or application, they search from the wide service pool by typing their desired service description. With this project, it is aimed to provide the services they need in real time when they are online by predicting them with artificial intelligence supported algorithms. The related topic is modeled as “Recommendation Engine” under the machine learning discipline. All service requests coming in 2020 were used as a training set. Since the queues of the services requested by the customers are interconnected in terms of temporality, the requested service queues are modeled according to the "Conditional Probability Based Prediction" method. In order to capture exceptional customer behaviors, customer specific habits have also been added to the service list. All machine learning models run on the AWS cloud ecosystem. It has been developed with the principle of running web services in Docker containers, which is the industry standard and used during the service of machine learning models to the customer. “Top-8 Service Accuracy” was chosen as the success metric of the project. The success rate of 22%, which is currently achieved by combining popular services throughout Turkey, has been increased to 37% with the new algorithm supported by AI. This rate comes up to 44% when we look at the customers who have had at least 1 service request in the past. The current service recommendation system, which has difficulty in capturing special customer behaviors with the perspective of “popular service throughout Turkey”, has significantly improved with the new AI supported approach by taking into account customer habits and the relationship between services. In the next stages of the project, innovative methods used in this field such as “Collaborative Filtering” and “RNN” will be performed together with the “ClickStream” data of the customers and the success rate will be tried to be increased.

Anahtar Kelimeler

• Artificial Intelligence
• Recommendation Engine
• Conditional Probability Based Prediction

Kaynakça

1. Adomavicius, G. ve Tuzhilin, A. (2005). Toward the next generation of recommender systems: a survey of the state-of-the-art and possible extensions. IEEE Transactions on Knowledge and Data Engineering, 17(6), 734-749. doi:10.1109/TKDE.2005.99
2. Ahmed, A., Kanagal, B., Pandey, S., Josifovski, V., Pueyo, L. G. ve Yuan, J. (2013). Latent factor models with additive and hierarchically-smoothed user preferences. Proceedings of the sixth ACM international conference on Web search and data mining - WSDM ’13 içinde (s. 385). the sixth ACM international conference, sunulmuş bildiri, Rome, Italy: ACM Press. doi:10.1145/2433396.2433445
3. Burke, R. (2002). Hybrid Recommender Systems: Survey and Experiments. User Modeling and User- Adapted Interaction, 12(4), 331-370. doi:10.1023/A:1021240730564
4. Herlocker, J. L., Konstan, J. A., Terveen, L. G. ve Riedl, J. T. (2004). Evaluating collaborative filtering recommender systems. ACM Transactions on Information Systems, 22(1), 5-53. doi:10.1145/963770.963772
5. Liu, J., Dolan, P. ve Pedersen, E. R. (2010). Personalized news recommendation based on click behavior. Proceedings of the 15th international conference on Intelligent user interfaces - IUI ’10 içinde (s. 31). the 15th international conference, sunulmuş bildiri, Hong Kong, China: ACM Press. doi:10.1145/1719970.1719976
6. Portugal, I., Alencar, P. ve Cowan, D. (2018). The use of machine learning algorithms in recommender systems: A systematic review. Expert Systems with Applications, 97, 205-227. doi:10.1016/j.eswa.2017.12.020
7. Rodriguez, M., Posse, C. ve Zhang, E. (2012). Multiple objective optimization in recommender systems. Proceedings of the sixth ACM conference on Recommender systems - RecSys ’12 içinde (s. 11). the sixth ACM conference, sunulmuş bildiri, Dublin, Ireland: ACM Press. doi:10.1145/2365952.2365961
8. Steck, H. (2013). Evaluation of recommendations: rating-prediction and ranking. Proceedings of the 7th ACM conference on Recommender systems içinde (ss. 213-220). RecSys ’13: Seventh ACM Conference on Recommender Systems, sunulmuş bildiri, Hong Kong China: ACM. doi:10.1145/2507157.2507160

9. Van Meteren, R. ve Van Someren, M. (2000). Using content-based filtering for recommendation. Proceedings of the machine learning in the new information age: MLnet/ECML2000 workshop, 30, 47-56. http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.25.5743 adresinden erişildi.
10. Zhang, C. ve Zhang, S. (2002). Association rule mining: models and algorithms. Lecture notes in computer science. Berlin New York: Springer.
11. Zhang, S., Yao, L., Sun, A. ve Tay, Y. (2019). Deep Learning Based Recommender System: A Survey and New Perspectives. ACM Computing Surveys, 52(1), 1-38. doi:10.1145/3285029