Diy Platform App Store Index Churn Analysis
Churn Analysis application predicts the customer churn score that indicates the probability of customer getting churned. Churn analysis is modeled as a typical classification problem. Training data is generated from the historical sales and payment data, a target label is assigned to each customer indicating whether he has churned or not. This training data is used to train a model which will be used to predict the churn score of new customers.
How to build my first Churn Analysis app?
To get you started immediately, we have provided a sample training data set of telecommunication business, which contains 19 different features and a target label named Churn. The value of target label is either True or False, defining whether the customer has churned or not.
To prepare your own training dataset Visit here
- Go to the app store, build a churn analysis app by searching for Churn Analysis
- You will be redirected to the app settings page where you can configure and run the app
- Under Quick Settings,
3.1. Select an existing dataset or upload a new dataset and click Save
3.2. Select the target field, the field that indicates whether customer is churned or not, and click Save
- Now click on Run Now to start the app. The app may run for 5 to 20 minutes (dependending on the size of input data) and then the trained model will be saved.
- Use this model to get the churn prediction
5.1. Click on Build App on the model to build the inference app using that model
5.2. Provide the input in the given fields and click on Get Results. The predicted churn class and scores will be displayed on the right side of the screen.
- [Optional] To use this model to predict more than one examples, [Visit here] [TODO]
The output of the infer app will the predicted class (churn or non churn) and probability score of prediction.
Go ahead and build your first Churn Analysis App