MyDataSay Android App for Predictive Model Deployment / Download

MyDataSay is an Android App for deploying predictive models on Android devices such as phones and tablets. It provides a simple easy way to deploy your predictive models for sales people who need heuristic information from your past data. Note that predictive models are developed using CMSR data miner from your historical data such as insurance claims data, credit delinquency/default records, etc.


To download the App, click MyDataSay Download. It's a .zip file. Read the "USER-INSTALL-GUIDE.txt" for installation on your Android devices. Android 3.0.0 or newer versions are supported. Web-server version is also available. If you are interested in web-server version, please write us.


The following screenshots show how end users will use MyDataSay Android App;

(Example Screenshots)


1. User selects a predictive model from the following screen, say,
"Customer churn risk", by touching/clicking it;
MyDataSay Android App: Selecting a model.



2. User enters model input data as in the following screen and
click the "Predict" button;
MyDataSay Android App: Selecting a model.



3.Then the model shows the following result/number;
MyDataSay Android App: Selecting a model.



4. User clicks "About Model" to assess the prediction.
The following information is displayed. It provides information how
the result can be interpreted;
MyDataSay Android App: Selecting a model.


5. User clicks "Visualize" to understand how the prediction is made.
The following screen show neural network internal status. It shows how
the result is produced;
MyDataSay Android App: Selecting a model.


The following screen is decision tree internal status. It shows how
the result is produced in decision tree;
MyDataSay Android App: Selecting a model.



To download the App, click MyDataSay Download. It's a .zip file. Read the "USER-INSTALL-GUIDE.txt" for installation on your Android devices. Android 3.0.0 or newer versions are supported. Web-server version is also available. If you are interested in web-server version, please write us.