liuzy1992 / sentclassifier

Do sentence classification using several machine learning methods.

Description:

sentclassifier is used to classify sentences of PubMed papers based on machine learing models.

Usage:

    ./sentclassifier.sh <-i str> [-m str] [-n int]

Parameters:

-i: input file with a header in following TSV format:
pmidsentidsenttextlabel

-m: machine learning model to use. choose from ['svm', 'rf', 'nb', 'knn']:
'svm': Support Vector Machine (Default)
'rf': Random Forest
'nb': Multinomial Naive Bayes
'knn': K-NearestNeighbor

-o: directory name for saving trained model. default is previous directory.

-n: number of threads to be used. default=1

Example:

Use following command to run an example:

    ./sentclassifier.sh -i testdata/test.tsv -m svm -n 1