Please use this identifier to cite or link to this item: http://localhost:8080/xmlui/handle/123456789/1742
Title: A Comparative Study of Statistical Analysis on Big Mart using Data Mining Techniques
Authors: T K Thivakaran
Dr M Ramesh
Issue Date: 2021
Publisher: The World Academy of Research in Science and Engineering
Citation: International Journal of Advanced Trends in Computer Science and Engineering
Abstract: In order to estimate sales revenue that is tangible and achievable, businesses involved in wholesales, manufacturing activities, marketing activities, retailing, logistics and supply chain activities need to use historical transaction data to forecast sales. In order to do this, there are several traditional data mining and statistical techniques that are used to identify trends, make predictive as well as descriptive analysis. The knowledge gained from such analysis is used in making business decisions. The data set in this study has been collected in the year 2013, and has 1559 products across 10 stores in different cities. First we conduct Exploratory Data Analysis to understand the nature of thAfter this, several traditional and novel data mining techniques have been applied on this data set, namely, linear regression, ridge regression, random forest regressor, decision tree regressor, XG Boost regressor and ARIMA. The cross-validation scores of all models are compared and inference as to which attributes and feature are given most weight during prediction of Item Outlet Sales attribute (target attribute) in the data set. Towards the end of the paper, the inferences and results are noted and discussed, hence completing the entire data analysis cycle.
URI: http://localhost:8080/xmlui/handle/123456789/1742
Appears in Collections:Computer Science Engineering Department

Files in This Item:
File SizeFormat 
SOE-CSE-16.pdf141.95 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.