This project/internship titled "Market Basket Analysis Approach to Machine Learning" submitted by Md Abul Hasnat Patwary, ID No Md Tamim Eshan, ID No Prazzal Debnath, ID No at Department of Computer Science and Engineering, International University Daffodil has been accepted as satisfactory in partial fulfillment of the requirements for the degree of B.Sc. Department of Computer Science and Engineering Faculty of Information Science and Technology Daffodil International University. We declare that this thesis has been done by us under the supervision of Abdus Sattar, Assistant Professor, Department of CSE Daffodil International University.
We are grateful and indebted to Abdus Sattar, Assistant Professor, Department of CSE Daffodil International University, Dhaka. Akhter Hossain, Professor and Head of CSE Department, for his kind assistance in completing our dissertation and also to other faculty members and staff of CSE Department of Daffodil International University. We would like to thank our entire coursemate at Daffodil International University who participated in this discussion while completing the course.
Here is a discussion of their methods so that traders in Bangladesh can make the right investments at the right place. Frequent sets of objects are retrieved from the database using the Apriori algorithm, and then association rules are generated. In this theory, the buyer tries to extract from the purchased object the relationship of another object. The main objective of market basket analysis (MBA) in field marketing, nuclear science, etc. is to provide data to distributors.
Understand the buyer's purchasing behavior, which can help the retailer make the right decision.
Motivation
Problem Definition
The data mining program will be able to state with high confidence what is selling well with the sticker with only one or two sticker customers, but it may only be correct for one or two valued customers. If cola discount coupons are already placed on frozen pizza from the convenience store, the fact that cola and frozen pizza sell well together will not be a surprise to them - it does not provide any new details, it just shows that current strategies of marketing are already running. In reality, a true partnership can even be obscured by the previous campaign - maybe people would normally prefer to buy beer with pizza, but because of the discount, they just buy Coke.
The general store is losing out on what could be a better promotion in this situation.
Objective
Research Question
Research Question 4: What kind of algorithms and what technology can we use to use this model.
Expected Outcome
Layout of the Report
This study aims to determine the effects of implementing consumer basket analysis in a retail store. We can define in advance that the development and evaluation of the model is the most time spent on the project. Most of the time, there is a general misunderstanding about what machine learning is and its capabilities.
Data extraction is the first step that can be considered part of the process of data transformation. This cleaning method is one of the most distinct aspects of a data science project done at university and in the real world. Most of the project time is spent on this task due to these high costs.
It is the features that add value to the model, rather than the hyperparameter configuration of the algorithm. To capture the evolving behavior of the data, these machine learning models must be retrained periodically. This term of the problem must be specified with the client and according to the requirements.
There are several metrics to measure the efficiency of a model, depending on the type of problem. Daffodil International University 17 Focus on the denominator, it is the probability of the individual support values of A and B and not together. All item groups except E have a support number greater than or equal to the minimum support, so E will be removed.
2. The join and prune algorithm steps can be easily implemented on large datasets. In the same way, you have to take care of the traditional rituals of the people in every society. Because discounting the usual common stuff doesn't make that much of a difference in the amount of product sold.
LITERATURE REVIEW ............................................................... 06-10
Related Works
They built a, Data cleaning method helps improve the consistency of the input data set and hence the results of the MBA by removing all kinds of errors from it. Using the neural network approach, the current Apriori algorithm is updated to optimize the prediction effects. Cleanup, as none of the current approaches considered the possibility of raw data or noisy data in Via's purchase history.
They are unaware of the customer's shopping preferences because they do not know what merchandise should be assembled in their store. Personalized reviews have also emerged as part of the marketing process since the rise of e-commerce. As a result, if you know in advance which goods are usually bought together, it is possible to adjust the distribution of the store to sell more products and the last domain they are talking about, to establish offers and promotions.
Most of the time, the data sets used for learning and practice are already clean and do not need to be handled. Part of the job is to know what information is important or can provide value to the algorithm and consider it on a case-by-case basis. The process of using domain knowledge of data to create features that enable machine learning algorithms to operate is feature engineering.
The approach involved the development, transformation and deletion of features, and the consistency of the model generated with these features was checked in all cases. To create connection rules, we will start by creating a new table with all possible rules from the combination of A, B.C. Now through the box plot we can know which of these products has more trust and support.
It should also be noted that when you have a department store in an area, you have to take care of the economy, tradition, holidays, etc. It should be made in a good residential area and the price of the product should be commensurate with the quality of the product so that they are interested in buying it. But if you notice that people with hamburgers like to eat sandwiches, but they can't take them or don't care because of the price, you can keep discounts to increase sandwich sales.
Challenges
RESEARCH METHODOLOGY ............................................... 11-14
- Business Goals and Objectives
- Data Extraction
- Data Cleaning
- Feature Engineering
- Model Creation
- Model Evaluation
- Business Impact Analysis
- Introduction
- Association Rule Mining
- Apriori Algorithm
- Steps for Apriori Algorithm
- Apriori Algorithm Working
- Advantages of Apriori Algorithms
- Disadvantages of Apriori Algorithms
- Python Implementation of Apriori Algorithm
- Data Collection Procedure
- Data Format and Statistical Analysis
- Data Pre-Processing
- Training the Apriori Model on the dataset
- Data Visualization
- Output
Cleaning data is often used in the process of eliminating data that is not important or necessary. The success of it is the result of all the work done in the process. The second is to validate the model and evaluate its output in the actual data.
In Data, we can assist our customer to understand what data says about a business, but in the end, the customer is the one who has to implement the corresponding behavior. In the first step, we will create a table that contains the support statement (The frequency of each item set individually in the data set) of each item set in the given data set. We will get the auxiliary count from the main transaction table of datasets after creating the subsets, which is how many times these pairs occurred together in the given dataset.
As seen in the C3 table above, there is only one item set combination with a support score equal to the minimum support score. Each row in the data set shows the goods customers bought or the transactions they made.
CONCLUSION AND FUTURE WORK
If you lower the price of cocoa now, there will be some profit in selling the product, as they will buy this product. Again, if you want, you can increase product sales by offering or discounting products that are not selling. But if you see Iggy Noodles being sold, then you can offer that if someone buys 5 packets of Iggy Noodles, they will get a Maggie Noodles for free.
Kang, “Market Basket Analysis: Identify the changing trends of market data,” in International Conference on Computational Modeling and Security (CMS 2016), Sangrur. Gore, "Optimized Predictive Model using Artificial Neural for Market Basket Analysis," Research Gate, Pune, Maharashtra, India, 2017. Wagner, "Using Market Basket Analysis to Integrate and Motivate Issues in Discrete Structures," in ACM SIGCSE Bulletin, Eau Claire , 2006.
