Market segmentation for profit maximization using machine learning algorithms

Abstract

In this present era with growing population, markets play an important role in providing the required and desired utilities to the people. For this it is important to enhance the Customer Relationship Management which can be achieved by segmenting the market utilities by various factors like weekly sales, demand and supply. In this research paper we discover the valuable information that is weekly sales and develop an efficient business strategy model to increase the profitability of the market through supply as well meet the demands of the customers. Our objectives are i) To find the top profitable products of the market across all the branches as well as their correlation with other products, ii) Understand the customer behavior according to the market flow and iii) Forecast the sales using ARIMA (Auto Regressive Moving Average). In this project, as we lack proper information about the customer identity, we use k means clustering which is an unsupervised learning model to cluster the customers according to the weekly sales behavior. Since we concentrate on the Weekly Sales to understand the market behavior, the best method to be applied on the dataset is the moving average technique. Here, we use ARIMA model to get the best results for forecasting as our time series data is considered to be stationary.