# kaefa kwangwoon automated exploratory factor analysis for improving research capability to identify unexplained factor structure with complexly cross-classified multilevel structured data in R environment

Using machine learning techniques with Kepler Space Telescope exoplanet search data to train and tune classification models

Predicting chess game outcomes using logistic regression and Elastic Net on 117k+ Lichess games. Includes full data processing, modeling pipeline, and interpretation of key predictors.

Tools to perform model selection alongside estimation under Linear, Logistic, Negative binomial, Quantile, and Skew-Normal regression. Under the spike-and-slab method, a probability for each possible model is estimated with the posterior mean, credibility interval, and standard deviation of coefficients and parameters under the most probable model.

森林総合研究所北海道支所所内セミナーでの発表資料「いつでもAICを使えばよいというものではない」

This GitHub repository features R code for a movie rating prediction project. It utilizes tidyverse in R to clean data, perform linear regression modeling, and evaluate prediction accuracy using RMSE and MAE metrics. The project aims to predict IMDB scores based on movie features such as gross revenue, budget, release year, and content rating.

Comparison of model selection methods for Boston dataset

In this project we can see in action and in detail a big part of the ML pipeline (data wrangling,model building, model evaluation) that comprises different algorithms and approaches such as Decision Trees (RPART), Linear Discriminant Analysis (LDA), Gradient Boosting Machne (GBM), Random Forest (RF) Support Vector Machine (SVM) with or without M…

A predictive analytics engine for facial emotion recognition comparing multiple machine learning strategies, including XGBoost, LDA, KNN and Random Forest, using facial fiducial point distributions.

Various variable selection methods are explored

This project explores linear regression model selection in R using Best Subset Selection (BIC), stepwise methods with cross-validation, Ridge, and Lasso. Includes MSE evaluation on test data, multicollinearity analysis (VIF), and correlation insights for variable selection.

Build predictive models to identify students at risk of dropping out, using their demographic, academic, and social indicators.

Optimal topic identification from a pool of Latent Dirichlet Allocation models

Predict number of students on private and public school in Hawaii

This project was in collaboration with University Hospital Birmingham

R programming scripts for statistical data analysis and visualization. Includes data cleaning, descriptive statistics, hypothesis testing (t-tests, ANOVA), correlation, regression modeling, and visualizations using ggplot2 and tidyverse.

Generación de reportes para los mejores modelos Logit obtenidos con el paquete glmulti

We created an exploratory model to identify if the given marketing campaign helped increase the revenue

Determined the best regression model which represents the data