A Novel Prediction of Sales and Purchase Forecasting for Festival Season of Hypermarkets with Customer Dataset Using Apriori Algorithm Instead of FP-Growth Algorithm to Improve the Accuracy

Abstract

Aim: To predict the novel and to forecast sales for festival season hypermarkets. Materials and Methods: A total of 484 samples were collected from market datasets available in kaggle. For this two algorithms were used, one is the FP-Growth algorithm and another is Apriori algorithm. Both the algorithms were executed and compared for accuracy. Result: Apriori achieved accuracy, precision, sensitivity and specificity of 73 %,75%, 78%,and 80%, respectively, compared to 71%, 73%, 76%, 75%, and 78% by FP-Growth algorithm, 87.4%, 88.2%, 89.2%, and 93%, respectively, compared to 80.1%, 83.39%, 84%, and 86.20% by Apriori algorithm. The results were obtained with a level of significance (p<=0.310). Conclusion: The applied Apriori algorithm confirms to have higher accuracy than the FP-Growth algorithm. It was additionally found that FP-Growth calculation takes more modest time than Apriori calculation to yield novel results.