1230005536191_pThe dataset used in this project is a transnational dataset which contains all the transactions occurring between 01/12/2010 and 09/12/2011 for a UK-based and registered non-store online retail. The company mainly sells unique all-occasion gifts. Many customers of the company are wholesalers. You will be using the online retail transnational dataset to build a RFM clustering and choose the best set of customers which the company should target./ppIn this project, you will perform Cohort analysis and RFM analysis. You will also perform clustering using K-Means to get 5 clusters. The machine learning models used in this project to predict clusters as target variable are K-Nearest Neighbor, Random Forest, Naive Bayes, Logistic Regression, Decision Tree, Support Vector Machine, LGBM, Gradient Boosting, XGB, and MLP. Finally, you will plot boundary decision, distribution of features, feature importance, cross validation score, and predicted values versus true values, confusion matrix, learning curve, performance of the model, scalability of the model, training loss, and training accuracy./p(*_*)1230005536191_

In this project, we embarked on a comprehensive journey of exploring the dataset and conducting analysis and predictions in the context of online retail. We began by examining the dataset and performing RFM (Recency, Frequency, Monetary Value) analysis, which allowed us to gain valuable insights into customer purchase behavior.

Using the RFM analysis results, we applied K-means clustering, a popular unsupervised machine learning algorithm, to group customers into distinct clusters based on their RFM values. This clustering approach helped us identify different customer segments within the online retail dataset.

After successfully clustering the customers, we proceeded to predict the clusters for new customer data. To achieve this, we trained various machine learning models, including logistic regression, support vector machines (SVM), K-nearest neighbors (KNN), decision trees, random forests, gradient boosting, naive Bayes, extreme gradient boosting, light gradient boosting, and multi-layer perceptron. These models were trained on the RFM features and the corresponding customer clusters.

To evaluate the performance of the trained models, we employed a range of metrics such as accuracy, recall, precision, and F1 score. Additionally, we generated classification reports to gain a comprehensive understanding of the models predictive capabilities.

In order to provide a user-friendly and interactive experience, we developed a graphical user interface (GUI) using PyQt. The GUI allowed users to input customer information and obtain real-time predictions of the customer clusters using the trained machine learning models. This made it convenient for users to explore and analyze the clustering results. The GUI incorporated visualizations such as decision boundaries, which provided a clear representation of how the clusters were separated based on the RFM features. These visualizations enhanced the interpretation of the clustering results and facilitated better decision-making.

To ensure the availability of the trained models for future use, we implemented model persistence by saving the trained models using the joblib library. This allowed us to load the models directly from the saved files without the need for retraining, thus saving time and resources. In addition to the real-time predictions, the GUI showcased performance evaluation metrics such as accuracy, recall, precision, and F1 score. This provided users with a comprehensive assessment of the models performance and helped them gauge the reliability of the predictions.

To delve deeper into the behavior and characteristics of the models, we conducted learning curve analysis, scalability analysis, and performance curve analysis. These analyses shed light on the models learning capabilities, their performance with varying data sizes, and their overall effectiveness in making accurate predictions. The entire process from dataset exploration to RFM analysis, clustering, model training, GUI development, and real-time predictions was carried out seamlessly, leveraging the power of Python and its machine learning libraries. This approach allowed us to gain valuable insights into customer segmentation and predictive modeling in the online retail domain.

By combining data analysis, clustering, machine learning, and GUI development, we were able to provide a comprehensive solution for online retail businesses seeking to understand their customers better and make data-driven decisions. The developed system offered an intuitive interface and accurate predictions, paving the way for enhanced customer segmentation and targeted marketing strategies. Overall, this project demonstrated the effectiveness of integrating machine learning techniques with graphical user interfaces to provide a user-friendly and interactive platform for analyzing and predicting customer clusters in the online retail industry.

...

ONLINE RETAIL CLUSTERING AND PREDICTION USING MACHINE LEARNING WITH PYTHON GUI

Comentarios