Pinned Repositories
Artificial-Neural-Network-on-Churn-Modeling-Dataset-for-a-Bank
Artificial Neural Network on Churn Modeling Dataset, built from scratch using Keras in Python. This Helps a bank to predict whether a particular customer would be leaving the bank in the future or not(Binary Classification)
DiCE
Generate Diverse Counterfactual Explanations for any machine learning model.
DiCE--Counterfactual-Explanations-for-SBA-Loan-Default
dowhy
DoWhy is a Python library for causal inference that supports explicit modeling and testing of causal assumptions. DoWhy is based on a unified language for causal inference, combining causal graphical models and potential outcomes frameworks.
DoWhy-The-Causal-Story-Behind-Hotel-Booking-Cancellations
Machine-Learning-for-Thermoelectrics-Discovery
Transition metal oxides are attractive materials for high temperature thermoelectric applications due to their thermal stability, low cost bulk processing and natural abundance. Notwithstanding the high power factor, their high thermal conductivity is a roadblock in achieving higher efficiency. The search space for new thermoelectric oxides has been limited to the alloys of a few previously explored systems, such as ZnO, SrTiO3 and CaMnO3. The phenomenon of thermal conduction in crystalline alloys and its dependence on crystal properties is also poorly understood, which limits the ability to design new alloys. In this paper, we apply machine-learning models for discovering novel transition metal oxides with low lattice thermal conductivity (kL). A two-step process is proposed to address the problem of small datasets frequently encountered in materials informatics. First, a gradient boosted tree classifier is learnt to categorize unknown compounds into three categories of thermal conductivity: Low, Medium, and High. In the second step, we fit regression models on the targeted class (i.e. low kL) to estimate kL with an R2 value of 0.96. Gradient boosted tree model was also used to identify key material properties influencing classification of kL, namely lattice energy per atom, atom density, electronic energy band gap, mass density, and ratio of oxygen by transition metal atoms. Only fundamental materials properties describing the crystal symmetry, compound chemistry and interatomic bonding were used in the classification process, which can be readily used as selection parameters. The proposed two-step process addresses the problem of small datasets and improves the predictive accuracy.
MAT-494-Deep-Learning-SNU
Lab sessions taught by me for MAT-494-Deep Learning (Monsoon-2020) at Shiv Nadar University, India.
Network_Learning_approaches_to_study_World_Happiness
The United Nations in its 2011 resolution declared the pursuit of happiness a fundamental human goal and proposed public and economic policies centered around happiness. In this paper we used 2 types of computational strategies viz. \textit{Predictive Modelling} and \textit{Bayesian Networks(BNs)} to model the processed historical happiness index data of 156 nations published by UN since 2012. We attacked the problem of prediction using GRNNs and show that it out performs other state of the art predictive models. To understand causal links amongst key features that have been proven to have a significant impact on world happiness, we first used a manual discretization scheme to discretize continuous variables into 3 levels viz. \textit{Low, Medium} and \textit{High}. A consensus World Happiness BN structure was then fixed after amalgamating information by learning 10000 different BNs using bootstrapping. Lastly, exact inference through conditional probability queries was used on this BN to unravel interesting relationships among the important features affecting happiness which would be useful in policy making.
Predicting_Product_Sales_through_ads_on_SocialMedia_using_k-N.N.
We build a k-N.N. model to tell whether a user on Social Networking site after clicking the ad's displayed on the website,end's up buying the product or not.
Predicting_Startup_Profits_using_Multiple_Linear_Regression
This contains the model of how we can predict the profit obtained by a Startup based on certain features, making use of Multiple Linear Regression in Python.
Sid-darthvader's Repositories
Sid-darthvader/Artificial-Neural-Network-on-Churn-Modeling-Dataset-for-a-Bank
Artificial Neural Network on Churn Modeling Dataset, built from scratch using Keras in Python. This Helps a bank to predict whether a particular customer would be leaving the bank in the future or not(Binary Classification)
Sid-darthvader/DoWhy-The-Causal-Story-Behind-Hotel-Booking-Cancellations
Sid-darthvader/MAT-494-Deep-Learning-SNU
Lab sessions taught by me for MAT-494-Deep Learning (Monsoon-2020) at Shiv Nadar University, India.
Sid-darthvader/Machine-Learning-for-Thermoelectrics-Discovery
Transition metal oxides are attractive materials for high temperature thermoelectric applications due to their thermal stability, low cost bulk processing and natural abundance. Notwithstanding the high power factor, their high thermal conductivity is a roadblock in achieving higher efficiency. The search space for new thermoelectric oxides has been limited to the alloys of a few previously explored systems, such as ZnO, SrTiO3 and CaMnO3. The phenomenon of thermal conduction in crystalline alloys and its dependence on crystal properties is also poorly understood, which limits the ability to design new alloys. In this paper, we apply machine-learning models for discovering novel transition metal oxides with low lattice thermal conductivity (kL). A two-step process is proposed to address the problem of small datasets frequently encountered in materials informatics. First, a gradient boosted tree classifier is learnt to categorize unknown compounds into three categories of thermal conductivity: Low, Medium, and High. In the second step, we fit regression models on the targeted class (i.e. low kL) to estimate kL with an R2 value of 0.96. Gradient boosted tree model was also used to identify key material properties influencing classification of kL, namely lattice energy per atom, atom density, electronic energy band gap, mass density, and ratio of oxygen by transition metal atoms. Only fundamental materials properties describing the crystal symmetry, compound chemistry and interatomic bonding were used in the classification process, which can be readily used as selection parameters. The proposed two-step process addresses the problem of small datasets and improves the predictive accuracy.
Sid-darthvader/Predicting_Product_Sales_through_ads_on_SocialMedia_using_k-N.N.
We build a k-N.N. model to tell whether a user on Social Networking site after clicking the ad's displayed on the website,end's up buying the product or not.
Sid-darthvader/Predicting_Startup_Profits_using_Multiple_Linear_Regression
This contains the model of how we can predict the profit obtained by a Startup based on certain features, making use of Multiple Linear Regression in Python.
Sid-darthvader/Network_Learning_approaches_to_study_World_Happiness
The United Nations in its 2011 resolution declared the pursuit of happiness a fundamental human goal and proposed public and economic policies centered around happiness. In this paper we used 2 types of computational strategies viz. \textit{Predictive Modelling} and \textit{Bayesian Networks(BNs)} to model the processed historical happiness index data of 156 nations published by UN since 2012. We attacked the problem of prediction using GRNNs and show that it out performs other state of the art predictive models. To understand causal links amongst key features that have been proven to have a significant impact on world happiness, we first used a manual discretization scheme to discretize continuous variables into 3 levels viz. \textit{Low, Medium} and \textit{High}. A consensus World Happiness BN structure was then fixed after amalgamating information by learning 10000 different BNs using bootstrapping. Lastly, exact inference through conditional probability queries was used on this BN to unravel interesting relationships among the important features affecting happiness which would be useful in policy making.
Sid-darthvader/DiCE
Generate Diverse Counterfactual Explanations for any machine learning model.
Sid-darthvader/DiCE--Counterfactual-Explanations-for-SBA-Loan-Default
Sid-darthvader/dowhy
DoWhy is a Python library for causal inference that supports explicit modeling and testing of causal assumptions. DoWhy is based on a unified language for causal inference, combining causal graphical models and potential outcomes frameworks.
Sid-darthvader/Emozers
This contains the complete code for Emozers an Emotion Analyzer and Music Player based on Machine Learning and Techniques of Applied Probability(Randomized Algortihms).This was created for HackData 2017 at Shiv Nadar University and was the winner of 2nd prize in Dot Tech Category.
Sid-darthvader/Forecasting-Delhi-AQI
Sid-darthvader/go
The Open Source Data Science Masters
Sid-darthvader/Happiness-Loneliness-Analytics
Sid-darthvader/Luxembourg-Co2-Emissions
A real world ML model built on the Dataset by provided Société Nationale de Circulation Automobile which tells us about two key characteristics of motor vehicles: fuel consumption and carbon dioxide.
Sid-darthvader/Market_Mix_Models_Intro
Sid-darthvader/ML-for-thermoelectrics-discovery
Associated code to reproduce results for the paper "Machine Learning approaches to Identify and Design Low Thermal Conductivity Oxide Alloys for Thermoelectric Applications"
Sid-darthvader/Probability-Simulations
Under this project I would be writing codes(In Java) to simulate various real life events and generate the randomized output by a computer as if these tasks were being performed by a human.
Sid-darthvader/sid-darthvader.github.io
Github Pages template for academic personal websites, forked from mmistakes/minimal-mistakes
Sid-darthvader/Sid-s-workshop
GitHub Demo
Sid-darthvader/sids_page
Bootstrap themed Website using Github pages
Sid-darthvader/Tide_Receipt_Matching
Sid-darthvader/ud851-Sunshine