Regression analysis technique analyzes the interrelationships between different project variables that contributed to the project outcomes to improve performance on future projects.
In simple terms it is an Interpolation Technique used in statistics
Simple regression Equation is : Y = BX + C
- Y is dependent variable
- B -> Slope
- X-> Independent variable
The slope of a regression line (B) represents the rate of change in Y as X changes. Because Y is dependent on X, the slope describes the predicted values of Y given X and C is a constant
It determines the extent to which a relationship exists between two variables. If the relationship is strong enough, one can then accurately predict the values of one variable based on the values of another using a simple linear formula.
Usage in Project Management:
As in regression analysis, the stronger the relationship is between the two variables, the greater the accuracy in predicting their relationship. Project managers can easily see the relationship between two variables by using a simple linear formula and plotting the results on a chart. And Used together with the other analytical tools, they can use information for the better decision-making.
- Decide the X and Y variables in your project that can be quantified and develop regression analysis
- R2 is the regression coefficient value should be close to 1, suggests good regression model or strong relationship
This technique is used in Close project Process Technique to analyze the interrelationships between different project variables that contributed to the project outcomes to improve performance on future projects.
- Quality Management Processes use this technique to understand to study and measure the factors that influence the output. As part of close project such data can be analyzed and used for root cause analysis, creating estimation models for future projects
In the above table the inspection results of software components 1-19 has been tabled. Row number 2 indicates the inspection rate in terms of lines of software code inspected per hour by the reviewer
Row number three indicates the results of the review in terms of number of defects the reviewer could find from the software program code. Now the test is to find is there a relationship between number of defects (independent variable ) and the rate of inspection ( dependent variable ). If the relationship is established, we can optimize the output by controlling the rate of inspection.
Now we shall draw a regression plot with these data points and the resulting graph is show below: this graph can be drawn using excel worksheet.
After inserting the graph you can select the data and ask the excel to add trend line. You can choose appropriate type based on the R2 value.
The R2 value here is 0.654. since the value of > 0.5, the relationship is predictable and can be used for furcating or estimating the number of defects for a given review rate. This is how we can create parametric estimation models from multiple historical data
Based on the actual data this model can be further tuned to make it current after every project
- Article reviewed and improved by Priyanshu Tyagi, 145th PMP® Preparation Workshop Batch
PMBOK® GUIDE SIXTH EDITION