Skip to content

Developing a Recommender System for Customers in a Scan and Go store using Apriori Algorithm in Market Basket Analysis

Notifications You must be signed in to change notification settings

Nishadleo94/Thesis

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

31 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Developing a Recommender System for Customers in a Scan and Go store using Apriori Algorithm in Market Basket Analysis

Abstract

The aim of the research is to study the possible impact a recommender system can bring about in increasing the revenue for a fully-automated Scan and Go store. The core of the research involves building a recommender system for Customers Using Apriori Algorithm in Market Basket Analysis. This data mining technique makes it possible to create an association between various products with the help of historical transactional data. Once a customer shops an item, the related products could be recommended to them.

Furthermore, the business aspect of introducing a recommender system is also studied. This involves how the revenue can be improved while at the same time satisfying the customers. Basically, the application of recommender system in a physical store is examined.

The secondary research involves the study of various published papers related to relevant topics such as Market Basket Analysis, Recommender systems, Various association rules, use of big data, etc. The knowledge obtained from these papers pave the way for the primary research analysis.

The primary research consists of executing the Market Basket Analysis in CRISP-DM methodology. Further, people reactions to a fully-automated Scan and Go store is also analysed using online survey. All the obtained results are visualised with the help of appropriate tools.

The question of how the recommender system improves the revenue of a company while making their customers satisfied is tried to answer by in-depth analysis of results obtained from the primary research and knowledge obtained from the secondary research

Results and Findings

Exploratory Analysis - Tableau

The Shiny Application

Sentiment Analysis of People towards "Amazon Go" (Scan and Go Store)

Conclusion

Considering the lack of availability of data such a search history, amount of time spend on a particular product ,that make the online recommender systems of amazon.com more efficient, the Apriori algorithm for association rule mining was observed as the best algorithm for carrying out the Market basket Analysis in a Scan and Go Store.

Introducing the recommender system into the Scan and Go stores can indeed satisfy both business as well as the customers at the same time. This should encourage the retail industry to implement the suggested model. The customers can use this Virtual Private Assistant (VPA) to make their shopping more convenient.

Releases

No releases published

Packages

No packages published

Languages