From 639f0a0c742150129c8e1fa6fb7d9324cb12bfe6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Alejandro=20Moreo=20Fern=C3=A1ndez?= Date: Thu, 15 Nov 2018 15:13:26 +0100 Subject: [PATCH] last updates --- src/classifier.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/classifier.py b/src/classifier.py index c23fc08..13cec32 100644 --- a/src/classifier.py +++ b/src/classifier.py @@ -8,6 +8,7 @@ from doc_representation import * # TODO: understand normalization # TODO: mendel hall # TODO: wrap into an Estimator +# TODO: check versions (numpy, scipy, sklearn) probability=True SVM = SVC @@ -20,7 +21,7 @@ if SVM is SVC: path = '../testi' -reader = LoadDocuments(split_documents=True, function_words_freq=True, tfidf=True, tfidf_feat_selection_ratio=0.1, split_policy=split_by_endline, normalize_features=True) +reader = LoadDocuments(split_documents=True, function_words_freq=True, tfidf=False, tfidf_feat_selection_ratio=0.1, split_policy=split_by_endline, normalize_features=True) Xtr,ytr,ep1,ep2 = reader.load(path) # learn a SVM