jhashanti / Machine-Learning-with-R

####Supervised learning #####Linear Regression

• Linear Regression
• Assumptions
• Variance Inflation Factor
• T Test
• F Test
• Adj. R Squared vs R Squared
• Mape
• Interpreting Actual vs Predicted Plot
• Interpreting Residual vs Predicted Plot

File:LinearRegression.R

Required packages: car, lmtest, ggplot2

Input parameters:

    c_path_in                       - path pointing to the input .csv file
    c_path_out                      - output folder path
    c_var_in_independent            - one ore more independent variable(s)
    c_var_in_dependent              - one dependent variable

Outputs:

    parameterEstimates (Data Frame) - Contains estimates of all the inpendent variables used for building the model &
                                      is exported as Estimates.csv to the location "c_path_out"
    modelStatistic (Data Frame)     - Contains various model statistics & is exported as ModelStatistic.csv to the location                                                 "c_path_out"
    durbinWatsonTest (List)         - Contains statistic used for testing autocorrelation
    goldfledQuantdtTest (List)      - Contains statistic used for testing homoscedasticity
    ActualPredicted.png file is exported to the location "c_path_out"
    ResidualPredicted.png file is exported to the location "c_path_out"

#####Logistic Regression

• Logistic Regression
• Chi-Square Test
• Model Evaluation Metrics(Gains Chart, Lift Chart, K-S Test, Confusion Matrix, ROC)
• Hosmer-Lemeshow Test

File: LogisticRegression.R

Required packages: car, ResourceSelection, ggplot2

Input parameters:

c_path_in                        - path pointing to the input .csv file
c_path_out                       - output folder path
c_var_in_independent             - one or more independent variables
c_var_in_dependent               - one binary dependent variable
x_val_event                      - level of dependent variable

Outputs:

parameterEstimate(Data Frame)    - Contains estimates of all the independent varaible used for building the model & is 							   exported as Estimate.csv file to the location "c_path_out"
modelStatistic(Data Frame)       - Contains various model ststistics & is exported as ModelStatistic.csv to the location 						   "c_path_out"
hosmerLemeshowTest(List)         - Contains statistic for testing goodness of fit
ks_out(Data Frame)               - For measuring the performance of the classification model
GainsChart.png is exported to the location "c_path_out"
LiftChart.png is exported to the location "c_path_out"

File: LogisticModelAnalysis.R

Input parameters:

c_path_out                       - output folder path
modelObj                         - Logistic Model Object

Description:

Contains function for generating Confusion Matrix for all the cut points between 0.01 & 0.99. ROC curve & Sensitivity - Specificity curve is plotted & exported as ROC.png & Sensitivity-Specificity.png respectively to the location "c_path_out"

#####Bayesian Belief Network

Understanding Bayesian Network using bnlearn - R Package, 1 Scoring

#####Naive Bayes

• Simple Explanation
• Bayesian Network vs Bayes Classifier

####Classification #####EM Clustering

• Expectation Maximization

####Ensemble #####Random Forest

• Random Forest
• Confidence Interval

####Recommender System Introduction

####Factor Analysis #####Mutual Information

• Mutual Information

#####Linear Discriminant Analysis

• Linear Discriminant Analysis
• LDA & QDA
• PCA vs LDA

File: Distance.R

Description:

Contains various method for calculating distance between two vectors