Correlation Analysis
Description
The handler detects interrelation between the series of the input data set based on the calculated correlation coefficients. It is used to access the supposed dependence of factors.
Let's perform the correlation analysis based on the small table that contains data on the sales number of four types of goods by dates.
Source table:
| Date | Spaghetti | Tomato paste | Pasta | Coffee |
|---|---|---|---|---|
| 02.09.17 | 10 | 20 | 15 | 25 |
| 03.09.17 | 12 | 22 | 12 | 26 |
| 04.09.17 | 14 | 25 | 9 | 26 |
| 05.09.17 | 13 | 24 | 10 | 25 |
| 06.09.17 | 14 | 25 | 9 | 24 |
| 07.09.17 | 14 | 25 | 9 | 23 |
| 08.09.17 | 12 | 21 | 12 | 24 |
| 09.09.17 | 10 | 18 | 14 | 23 |
| 10.09.17 | 16 | 24 | 9 | 22 |
| 11.09.17 | 13 | 21 | 9 | 23 |
| 12.09.17 | 17 | 25 | 7 | 25 |
Let's define correlation of the "Spaghetti" item of goods with other goods based on the Pearson coefficients (it is required to mark the "Spaghetti" field in "Set 1" and other goods in "Set 2" in the handler wizard).
Output table:
| Caption | Caption | Pearson coefficient |
|---|---|---|
| Spaghetti | Tomato paste | 0,83 |
| Spaghetti | Pasta | -0,93 |
| Spaghetti | Coffee | -0,12 |
As shown in the table, the sale series of the "Tomato paste" item of goods is distinguished by very high positive correlation, whereas the "Pasta" item of goods is characterized by the negative correlation. Thus, it is possible to draw a conclusion that the "Tomato paste" item of goods is a complement product, whereas the "Pasta" item of goods is a replacement product. The sale correlation of the "Coffee" item of goods with the "Spaghetti" item of goods is negative but the absolute correlation value is small, consequently, it is possible to draw a conclusion that there is no interrelation between sales of these items of goods.
Ports
Input
Input data source (data table).
Output
Output data set: the table that contains information on correlation between the fields. It has the following structure:- Required fields:
- Fields|Name: the name of the first field in the correlation pair.
- Fields|Caption: the caption of the first field in the correlation pair.
- Fields|Name: the name of the second field in the correlation pair.
- Fields|Caption: the caption of the second field in the correlation pair.
- There are the following fields the availability of which is set by a user:
- Pearson: the Pearson correlation coefficients.
- Correlation function extremum: extremums of cross-correlation function.
- Kendall's Tau-b: the Kendall's Tau-b rank correlation coefficient.
- Spearman: the Spearman's rank correlation coefficient.
- Required fields:
Wizard
It includes the list of checkboxes that enable to select coefficients for correlation estimation:
- Pearson correlation coefficient enables to determine the strength and direction of the linear relationship between two processes that occur simultaneously.
- Kendall's Tau-b coefficient: the rank correlation coefficient that is used to identify the quantitative relationship between variables if they can be ranked. It is recommended to use for categorical data.
- Extremum of cross-correlation function enables to calculate the maximum absolute value of the correlation coefficients of two processes calculated for all possible time shifts. It should be applied if it is required to determine the linear relationship between two processes, or parts of the processes that occur with a certain time lag.
- Spearman's rank correlation coefficient is another version of the rank correlation. Corresponding ranks, not numerical values, are used for numerical fields to estimate the connection strength. Therefore, the Spearman coefficient will be 1 or -1 for any monotone sequences.
It is possible to select the series in the table to analyze interconnection. For each field from "Set 1" selected with a checkbox, correlation coefficients with the fields selected with checkboxes in "Set 2" will be calculated.