Correlation analysis

Correlation analysis is a statistical method used to evaluate the strength and direction of the linear relationship between two quantitative variables. In dataset there are usually more then just two variables. However, the idea is, using the correlation analysis, to investigate the correlation between each input variable and the output (target) variable. Several types of correlation analyses exist, each with its own methodology and applications. The list of most common types of correlation analyses are:

1. Pearsons Correlation Analysis - Measures the linear relationship between two continuous variables. The correlation analysis assumes data is normally distribute. Values range from -1 to 1.
2. Spearman's Rank Correlation analysis - Measures the strength and direction of the association between two ranked variables. Does not assume a normal distribution. Useful for ordinal data or non-linear relationships.
3. Kendall's Thau - Measures the strength and direction of the association between two variables. Based on the ranks of the data rather than the data values.
4. Point-Biserial Correlation - Used when one variable is continuous and the other is dichotomous. Dichotomous means something that has only two possible values or categories (e.g. yes/no questions, gender classified as male/female, light switch that can be wither on or off)
5. Phi Coefficient - Used to measure the association between two binary variables. The correlation analysis is similar to Pearsons correlation but its specifically used for binary data.
6. Tetrachoric Correlation - Estimates the correlation between two dichotomous variables that are assumed to be derived from underlying continuous variables. USed when both variables are binary and are assumed to come from normally distributed variables.
7. Polychoric Correlation - Estimates the correlation between two ordinal variables. Assumes the oridinal variables are proxies for underlying continuous variables.
8. Biserial Correlation - Used when one variable is continuous and the other is dichotomous, but the dichotomy is artificial (e.g. passing/falling a test). Assumes the dichotomous variable is a cut-off oint of an underlying continuous variable.
9. Parital Corelation - Measures the relationship between two variables while controlling for the effect of one or more additional variables. Helps to understand the direct relationship between the two variables of interest.
10. Canonical Correlation - Measures the relationship between two sets of variables. Useful in multivariate statistical analysis to understand the association between two multivariate datasets.
11. Distance Correlation - Measure both linear and non-linear relationships between two variables. Does not require the relationship to be linear or even monotonic.
12. Rank-Biserial Correlation - Used when one variable is ordinal and the other is dichotomous. Suitable for situations where one variable is a ranked variable and the other is binary.

These 12 correlation methods vary in their assumptions, the types fo data they are suited for, and their sensitivity to the nature of the data relationships. Selecting the appropriate type of correlation analysis depends on the specific characteristics fo the variables involved and the nature of the relationship being studied.