I've now implemented the cohort level feature extraction.
Per CategoryType, DataItems are extracted per sample and inserted into DataItemMatrix (n_features x n_samples; a ConcurrentHashMap) using [DataSource,ItemType,Key] attributes from DataItem as the keys. Once all samples are processed, DataItemMatrix is written to a file.
Multithreading is based on TaskExecutor. I copied the code in SampleTask and CupAnalyzer.
Hey Charles,
I've now implemented the cohort level feature extraction.
Per
CategoryType
,DataItem
s are extracted per sample and inserted intoDataItemMatrix
(n_features x n_samples; aConcurrentHashMap
) using [DataSource
,ItemType
,Key
] attributes fromDataItem
as the keys. Once all samples are processed,DataItemMatrix
is written to a file.Multithreading is based on
TaskExecutor
. I copied the code inSampleTask
andCupAnalyzer
.