Establishing of a product recommendation system based on text similarity and singular value decomposition algorithm in a retail company

Due to rising client demand following the pandemic, organizations' initiatives in the field of digital transformation are gaining pace nowadays. Product recommendation systems were necessary for this context to deliver the most relevant product choices to clients, meet their expectations, and accomplish the businesses' sales targets through the e-Commerce website and mobile applications. Examining the product suggestions of numerous websites reveals that study on this subject frequently results in the presentation of illogical and erroneous product recommendations to visitors, given the variety and type of things accessible. This study looked into data manipulation, building custom functions, Text Similarity, and Singular Value Decomposition. The algorithm displays the best complementary commodities. The software was completed by developing the best algorithm. The product proposals were brought to life through the use of a website and a mobile application. Individual observations were made using Google Analytics and Python scripts. The study found that the created product recommendation system beat the current system by 7.62 per cent in items sold as a group and by 11.2 per cent in products sold separately. Furthermore, the number of views on the website and mobile application for the relevant area climbed by 7.17 per cent, and the number of clicks increased by 28.93 per cent. It has been decided that the product suggestions provide more logical and complementary product recommendations than the existing situation, and the relevant regions have been monitored and reviewed at various times.