Association Rules Mining for Season-Specific Time Frame in a Retail Supermarket

In the realm of data mining, analysis of the market basket is a common issue. Many associations have been found from the item sets of transactions after the daily transactions have been analyzed for different seasons or the whole of the year with a specific time frame. In a retail market, there are numerous products present in their database that are regularly bought just during a certain period in a day of a specific season, even though they are not often thought of as frequently purchased products. For example, commodities like bread and cake are more in demand during the morning session, whereas beverages and noodles are more in demand during the afternoon and evening sessions. Similarly, Hot coffee has a high demand in the winter season whereas in the summer, there is a huge demand for ice cream. During the monsoon season, cake and cookies are in high demand. Rice and soybean oil have high demand in the whole year. This study focused on identifying different types of frequent items as well as recognizing the connection among several frequent item sets by analyzing supermarket transactions. Besides this, the FP-Growth algorithm is utilized in our research to determine several relevant rules for different time frames of a day (morning, afternoon and evening) within a certain season (summer, monsoon and winter) and the entire year by analyzing the transactions. Our empirical research shows that the proposed method is computationally efficient at different times in a day of a specific season or whole of the year, helping a retail supermarket, a Bangladeshi retailer, make the right choices and increase sales.