association rule mining and regression analysis

blue: evaluation metrics

association rule mining

technique that discovers the strong associations or interesting correlations among given data in large transactional repositories.

if-then

ARM algorithms discover high-level prediction rules in the form:

antecedent

soda → chips, soda?

consequent

soda → chips, chips?

k

a _-itemset is an itemset with _ items

transaction t

a set of items and t ⊆ I

T

set of transactions

support

indicates how frequently the itemsets X. Y occur in the dataset

confidence

percentage of transactions that contain X also contain Y

lift

indicates how likely Y is satisfied when X is true, used to compare actual confidence compares to the expected confidence; indicates whether the probability of buying Y increases or decreases given X.

conviction

ratio of the probability that X appears without Y (if they were independent) with the actual frequency of the appearance of X without Y; used to measure the directional relationship between X and Y.

apriori algorithm

also known as level-wise search

regression

tool for building/developing a statistical (regression) model

simple regression

one independent variable

multiple regression

two or more independent variables

correlation analysis

degree and direction of relationship between variables

regression analysis

uses the relationship for predicting the value of a dependent variable or target variable

r-squared of coefficient of determination

represents the part of the variance of the dependent variable explained by the independent variables of the model; measures the strength of the relationship between the model and the dependent variable.

mean squared error (MSE)

the average of the squared difference between the predicted and actual value.

mean absolute error (MAE)

the average of the absolute difference between the target value and the value predicted by the model.

root mean square error

square root of the average of the squared difference of the predicted and actual value.

residual

difference between the predicted and actual value