Accuracy

A measure of how well one set of data matches another, particularly how well a set of predictions match the actual measures.  Accuracy is not a simple subject, it is not the opposite of error and there are many ways to calculate accuracy with widely different results.  The typical calculation of accuracy is 1-(abs(Predicted-Actual)/Actual), however this calculation breaks down if the Actual is zero (0).  Another example of an accuracy calculation is "Relative Accuracy" which is 1-(abs(Predicted-Actual)/(Range of Actual).

Causal Analytics

The analysis of data in an attempt to determine which variables cause each other, when and to what extent.

Causal Strength

The extent one variable causes another variable to move on a scale of 0-1, the sign of the value represents the direction.  A negative sign (-) means an inverse relation and a positive sign (+) means a direct relation.  Causal strengths less than 0.15 are somewhat likely to be random, less 0.10 is likely and 0.05 or less is quite likely to be random, as determined by analyzing cause and effect between random numbers.

Cleaned Data

Data that has been filtered to remove unwanted values.

Converted Data

Data that has been changed from one form to another, such as through value substitutions, application of mathematical functions, categorization, binning, normalization, etc.

Correlation

Common Definition: A measure of the similarity of movement of one data series to another. Often calculated using Pearson's method.

General Definition: The degree that one variable is related to another, the degree they are co-related.