- Find Data: 2566 studies founded
- First Filter: Review studies relevance: 12 studies
- Second Filter: Find common data schemas (ados exam) between the studies: 9 studies.
- Pre-proccessing each data set:
- Delete columns with more than 25% of NaN values.
- Remove duplicate columns
- Parse the columns name (features): All in lower case and removing special characters.
- Merge the 9 studies.
- Standardization of the diagnosis format (y values)