]Analysis: 1. level
COCO-STEP workflow
Preconditions for using COCO-STEP workflow:
- In case of price/performance analyses COCO STD and COC-S STD program executions may base further partial goals of analyses.
- If the first run of the similarity analysis gave balanced (error-free) results, so there is seemingly no winner of a price/performance analysis, but not all of the attributes were used to derive the balanced state, then further steps can be taken.
- In cases like this, multiple similarity analysis can be necessary to run in order to get a result that may be considered proven.
- The variant evaluation forces to take the available (that may become noise in some analyses) attributes into the analysis step by step.
- With the STEP IX analyses (multiple approximating steps) one can look for the minimal and maximal balanced prices evaluated by the each object.
- When forming forecasts and production functions the basic run can be done with any COCO-module.
- The simulation of decision trees may occur to us when we got a complex ceteris paribus relation as a result, but we suspect that the effects of different conditions are in the background. This can be also proved with multiple classification runs.
- A similarity analysis in itself is able to reveal contradictions, but there are some kind of contradictions that can be revealed only as the resultant of multiple partial results. For example: Out of the describing statistics of a region like labor, tax and regional development indicators, why would the regional development (drainage, tourism) indicator be explained with significantly less success than the others? (Maybe because it can be suspected to be the most biased decision making procedure?)
- Where did you use similarity analysis with steps based on each other before?
