PERFORMANCE EVALUATION OF PREDICT ALGORITHMS FOR ECOMMERCE, WITH FOCUS ON CUSTOMER SEGMENTATION

Customer segmentation refers to the classification of customers according to their attributes, behaviours, needs, preferences and values in a clear strategic business model and in a specific market. Dividing customers into different groups according to certain criteria and marketing them with targeted manners can improve customer loyalty and reduce customers’ lack of desire to patronize company’s services. Yang believes that customer segmentation refers to the classification of customers based on a combination of customer value, needs and interests in a clear strategic business model and focused market. There are multiple models (algorithms) that can provide a good predictive evaluation to a simple problem, and choosing the right one can be challenging. To combat this, one must understand the model’s performance evaluation by picking metrics that truly measure how well each model achieves the overall business goals of the company (In our case; Customer segmentation). Proper predictive performance models’ evaluation is also important because we want our model to have the same predictive evaluation across many different data sets. In other words, the results need to be comparable, measurable and reproducible, which are important factors that help the company make informed decision. Our project make use of a Mall customer data set, which contains 200 customer records, this datasets where performed on 3 predictive algorithms(K-means, Hierarchical, DBSCAN) and 5 performance metrics evaluate those models (Silhoutte, Rand-Index, Mutual information, Calinski-Harabasz Index, and DavisBouldin index). This project uses Scikit-Learn, which is a python library and an efficient tool for predictive data analysis. Link to dataset used and implementation code using jupyter notebook is made available for reference.