grigoryangayane Goto Github PK

feature_seleaction_importance_analysis

game-theory-genetic-algorithm-coalitions

This work utilizes inverse generative social sciences (a genetic algorithm), and agent-based modeling and simulation (ABMS) approaches, to dissect the dynamics of coalition formation across a spectrum of different coalition sizes illustrating the intricate patterns that emerge.

genetic_algorithm_abm

irl-maxent

Maximum Entropy and Maximum Causal Entropy Inverse Reinforcement Learning Implementation in Python

ml_edu

predator-prey-model-feature-importance

Simulation models encounter uncertainty, leading to significant output fluctuations from small input variations. To mitigate this, we consider a cooperative game theory-based feature importance method, which identifies uncertainty in datasets, aiding simulation model analysis. This approach enhances explainability of models with variable inputs.

red-wine-quality-prediction-using-regression-modeling-and-machine-learning

This project aims to determine which chemical features are the best quality red wine indicators. I used EDA, Regression modeling, LASSO, and Random Forest.

regressionshapley

Regression Shapley values are based on Lipovetsky and Conklin (2001) paper, which uses R-squared values to determine the feature contribution to the model performance. I have used the "seatpos" dataset to determine the feature importance when designing a car seat. The dataset has highly correlated variables, and the regression model results show statistically insignificant results. However, applying Regression Shapley Values reveals the more significant features for the prediction. I have conducted my analysis considering Explainable Artificial Intelligence (XAI). The paper can be found here: https://digitalcommons.odu.edu/msvcapstone/2022/scienceengineering/2/ References: Lipovetsky, S., & Conklin, M. (2001). Analysis of regression in game theory approach. Applied Stochastic Models in Business and Industry, 17(4), 319-330.

research_reproducability_analysis

This work focuses on leveraging explainable artificial intelligence (XAI) techniques to enhance the replicability and reliability of research findings in the social and behavioral sciences (SBS

software-agents-versus-humans-game-1

Game 1 refers to game with single core. The analysis compares rates of coalition formation between a special software agent versus an actual human player. The model is based on a solution concept from cooperative game theory called core and agent-based modeling approaches. The results of the analysis indicate some humans reach the core coalition faster in the earlier rounds when compared to the special software agent or the other computerized agents. This work has been published in Simulation: Transactions of the Society for Modeling and Simulation International journal.

software-agents-versus-humans-game-2

Game 2 looks at the coalition formation where for multiple core game case. The analysis compares rates of coalition formation between a special software agent versus an actual human player. The model employs a solution concept from cooperative game theory called core and agent-based modeling approaches. The results of the analysis have similar pattern as the Game 1 (single core coalition structure case). The results confirm some humans reach the core coalition faster in the earlier rounds when compared to the special software agent or the other computerized agents. This work has been published in Simulation: Transactions of the Society for Modeling and Simulation International journal.

tranportation_xai

AI applications can be found in various real-world systems, including vehicle system design and real-time car accident prediction. There is an increasing need to better explain AI-driven processes, especially in terms of potential legal disputes that might result from AI decisions. This analysis addresses this explainability and legal issues.

vgxai

New explainable artificial intelligence methods based on voting games cooperative game theory.

xai_cgt_methods