A graph-based recommendation system leveraging cosine similarity for enhanced marketing decisions

Abstract

PURPOSE: This work aims to present a comprehensive customer recommendation system based on cosine similarity. The primary objective is to develop an effective tool that assists sellers in identifying and recommending similar customers by analyzing their characteristics and behaviors.

DESIGN/METHODOLOGY/APPROACH: The methodology analyzes demographic data, purchase history, and other customer characteristics to calculate cosine similarity. This process includes data processing techniques such as feature integration and generating a cosine similarity matrix. The results demonstrate the system's effectiveness through thorough analysis.

FINDINGS: The analysis confirms the effectiveness of the proposed recommendation system, revealing that using cosine similarity can identify and recommend similar customers accurately.

PRACTICAL IMPLICATIONS: The study emphasizes incorporating modern data analysis methods into marketing and customer relationship management. This approach can enhance the efficiency of sales activities and elevate customer satisfaction.

ORIGINALITY/VALUE: This work offers a novel approach to customer recommendations by employing cosine similarity and innovative data processing techniques. It demonstrates how advanced data analysis methods can be leveraged to improve sales strategies and foster stronger customer relationships.