Unique Metric for Health Analysis with Optimization of Clustering Activity and Cross Comparison of Results from Different Approach. Paper Link
- A method for estimating clusters holistically. The metric dose not depend on any assumptions about the Approach to Cluster the Data. All that matters is the Data what it fits on and what are the Cluster Segments observed.
Reinforcement Evolutionary Learning Method for Self-Learning (RELM).Paper Link
- RELM is a take on using Reinforcement Learning on Quantitative Data where learning happens via Evolutionary Algorithms (GA, ES, NES, NSGA-II, etc). This method takes on the challenge age old problem of Concept Drift.
Futuristic Classification with Dynamic Reference Frame Strategy.Paper Link
- As a Data science practitioner, one of the classical problem is to understand The Churn. This paper takes a perspective on data that enables to identify what actually distinguishes Churn from Other.
Personalized Influence Estimation Technique(PIE). Paper Link
- We have lot of methods to access the global influence of a variable in data for most of ML Algorithms. But to make a justifiable point estimate about a datum might be a key source of information in many cases of Anomaly detection, Churn Analysis or Employee Attrition Reduction, etc.. Individual PIE values (point estimates) can be used to trigger a certain action. In this paper we estimate PIE values for Linear and Non Linear models.
Distribution Assertive Regression (DAR). Paper Link and Blog Post
- DAR is regression analysis in which we understand how a fit of regression varies over itβs value. In this paper we resolve the Regression by Parts and estimate the new data based on heuristic methods like K-nearest Neighbor and Regression Equation.