Integration algorithms of recommendation in the mobile trade system

Abstract

The algorithms of recommender systems must be efficient enough to process data without a significant load on the hardware of the devices. The solution to this problem includes optimization of calculations, minimization of memory usage and provision of quick response of the system to user requests. An important task is the integration and optimization of recommender systems in the field of mobile commerce using cloud services. The purpose of the work is a comparative analysis of different approaches to the in-tegration of recommendation algorithms in mobile commerce, with an emphasis on opti-mizing the user experience and increasing sales efficiency. In the software product, attention are paid to the design of a database for product categorization, a pricing system and information on retail outlets, the task of storing commercial documents, etc. Amazon Web Services was chose for dynamic business requirements due to its flexibility, scalability and comprehensive services. To create tasks with machine learning algorithms, the AWS Sage Maker service was used. When designing a recommender system, classification algorithms are using for accurate segmentation of customers. Time series analysis is key to understanding seasonal trends in sales, which helps predict future demand and inventory. Data dimensionality reduction techniques such as principal component analysis (PCA) can be used to efficiently visualize and extract insights from large data sets, facilitating interpretation and decision making. Batch analysis involves periodic processing of data. In the proposed solution, the recommendation system is the integration of selected services (Amazon S3, Amazon RDS, Amazon SageMaker, and Amazon Personalize) into the already existing mobile application infrastructure. The technology of practical implementation of machine learning algorithms in a mobile application for effective interaction with users and increasing their loyalty is considered. The integration of recommendation algorithms into mobile applications when using cloud services demonstrates the potential for further development in dynamic industries.