When you are in a rule, you have different indicators to explain the rule.
First, you have the main metrics of the rule.
- Population: how many points complied with the combination of settings (range selected for each dimension of the rule). Below, in "Out", you have the number of points of the objective variable that did not respect the rule (at least 1 of the conditions not applied).
- Presence: percentage of the analysis population that respects the rule. The information below, “Active dimensions” is the number of conditions activated in the rule.
- Result: average of the output variable for the points of the rule Population. Below, in "Out", you have the average of the objective variable that did not respect the rule (at least 1 of the conditions not applied).
- Lift: percentage of improvement inside the rule of the ratio of good vs bad. 0% = the same ratio than the study. 100% = we are always above our target set inside the analysis. In the green pie chart, you have the number of good points when the rule is respected.
- The chart shows:
- In blue, the average of the objective variable per period of time.
- In green, the average of the objective variable per period of time for the points that comply with the rule.
- The green bars show the presence per period of time.
You can click on the chart to have more details.
- The tab shows the statistics of the objective variable in the rule and in the analysis period.
You can find also indicators relative to the variables chosen for the rule.
- In the “Definition of the rule” frame, you are able to manage (add or remove) the variables of the rule (dimensions), modify the ranges and assess the impact on your output variable by clicking on “Recompute”.
- Lift impact: value of “Lift” that would be removed (if the variable is checked), respectively added (if the variable is not checked), if you remove (respectively add) the variable from the rule. This is given the other variables currently activated.
- If the lift of a variable is negative, it means the result of your objective variable will be increased if you add this variable to your rule.
- If the lift of a variable is positive, it means the result of your objective variable will be decreased if you add this variable to your rule.
- If the lift is 0, it means the result of your objective variable will be the same if you add this variable to your rule (the others variables have more impact).
- Population impact: same as above but for the number of points complying with the rule.
Remove variable A from the rule and 25 points will be added to the presence.
Remove variable B from the rule and 15 points will be added to the presence.
- Presence: percentage of the study population that respects the range chosen for the variable.
- Statistics:
- In the green column, statistics for this variable for subset of points complying with the rule.
- In white, statistics for this variable for all the points of the study.
- The chart represents the performance evaluation:
- The blue curve shows the grand average of the output variable and the green curve is the average of the variable inside the rule.
- The green bars show the application effort of the rule by time.
You can change the model of the chart: Histogram, Modal Histogram, Hyperlift, Presence by date.