Predikt : a simpler machine learning experience




Prakash, Piyush

Journal Title

Journal ISSN

Volume Title



With the exponential growth in volume of data being generated by the businesses today, the need for automated systems to facilitate decision making is growing. In many cases, the volume of data has reached the scale where it is not humanly possible to analyze the relevant data using traditional processes, which are dependent on manual intervention, to reach a good decision. The increase in diversity of data sources (internal, external, social, sensors, etc) has added significantly to the challenge of decision making. Predikt presents a simplified approach for the automation of decision making in organizations using Data Mining and Machine Learning techniques with minimal human intervention. Human Intervention, when required, will typically be to accept, reject or override a decision already made by the machine. Predikt front-end as well as back-end is able to adapt to end users unique requirements. Predikt is able to render User Interface dynamically based on User Inputs, while backend is able to do intelligent data processing such as binning (for numeric columns) and multiple model evaluation.



LCSH Subject Headings