- evaluateColumnValue(String) - Method in class org.canova.cli.csv.schema.CSVSchemaColumn
-
This method collects dataset statistics about the column that we'll
need later to
1.
- evaluateInputRecord(String) - Method in class org.canova.cli.csv.schema.CSVInputSchema
-
- evaluateStatistics() - Method in class org.canova.cli.transforms.image.NormalizeTransform
-
- evaluateStatistics() - Method in class org.canova.cli.transforms.text.nlp.TfidfTextVectorizerTransform
-
This is where we'll take the dataset stats learned from the first pass and setup for the
transform pass
- evaluateStatistics() - Method in interface org.canova.cli.transforms.Transform
-
- execute() - Method in interface org.canova.cli.subcommands.SubCommand
-
Execute the input
- execute() - Method in class org.canova.cli.subcommands.Vectorize
-
- execute() - Method in class org.canova.cli.vectorization.AudioVectorizationEngine
-
- execute() - Method in class org.canova.cli.vectorization.CSVVectorizationEngine
-
This is where our custom vectorization engine does its thing
- execute() - Method in class org.canova.cli.vectorization.ImageVectorizationEngine
-
In this case we are assuming that the Image input format gave us basically raw pixels
Thoughts
- Inside the vectorization engine is a great place to put a pluggable transformation system [ TODO: v2 ]
- example: MNIST binarization could be a pluggable transform
- example: custom thresholding on blocks of pixels
- execute() - Method in class org.canova.cli.vectorization.TextVectorizationEngine
-
Currently the stock input format / RR gives us a vector already converted
- TODO: separate this into a transform plugin
Thoughts
- Inside the vectorization engine is a great place to put a pluggable transformation system [ TODO: v2 ]
- example: MNIST binarization could be a pluggable transform
- example: custom thresholding on blocks of pixels
Text Pipeline specific stuff
- so right now the TF-IDF stuff has 2 major issues
1.
- execute() - Method in class org.canova.cli.vectorization.VectorizationEngine
-
- execute() - Method in class org.canova.cli.vectorization.VideoVectorizationEngine
-
In this case we are assuming that the Image input format gave us basically raw pixels
Thoughts
- Inside the vectorization engine is a great place to put a pluggable transformation system [ TODO: v2 ]
- example: MNIST binarization could be a pluggable transform
- example: custom thresholding on blocks of pixels