A Predictive Model in 3 steps…
Auto-ML Extension in Business Intelligence Platforms
Step 1
Using MachineWorks platform extension for BI tools, you can now leverage curated data sets within the BI platforms such as QlikSense and begin the model building pipeline. This was not possible before.
If you prefer to bring you own training dataset, simply enable the UseExternal toggle in the property sheet so that the extension allows you to browse and identify your external training file. Once uploaded, simply identify the desired fields from the BI tool and let the platform extension do the rest, starting with automated Exploratory Data Analysis to present you with descriptive statistics to better understand the distribution, central tendencies and variability in the data.
Our platform open architecture allows the BI extension to tap into the Machineworks powerful Jarvis engine to readily return uni-variate statistics in an intuitive interactive grid with built in visualization to show the distribution of the data. Users can very easily select, deselect features for model training as well as identify label for prediction.
Step 2
Once the predicted label identified, simply initiate the model training and let the MachineWorks platform do the rest of the automated model building pipeline for you. The platform uses multiple well-known Automated Machine Learning techniques for pipeline build and selection. Platform automatically identifies the problem type and chooses the appropriate algorithms to train using the training data. The extension also allows you to pick your own algorithm to train with, should you prefer to retain more control over the model selection. The best part of the experience is that you can iterate through this process as many times as you want, with as many permutations of the feature selection and refine the model until you are satisfied with the result, all without leaving the Qliksense BI platform.
Step 3
Now that the model is fully trained with your selected features and label, you are ready to generate predictions. Further, the platform the trained model is automatically deployed and is accessible for you to generate batch prediction or run real-time prediction using the API-KEY made available to you as the QlikSense variable. Simply upload your file to generate prediction and download it for further analysis.
Would love to hear your experience using our extension. Feel free to signup with MachineWorks and try our extension within your QlikSense BI Platform.
— The MachineWorks Team
