pipeline\u8fd9\u4e2a\u8bcd\uff0c\u5e94\u8be5\u6765\u81ealinux\u3002\u5728linux\u4f53\u7cfb\u4e0b\u7684\u5404\u79cd\u547d\u4ee4\u5de5\u5177\u7684\u5904\u7406\uff0c\u652f\u6301pipeline\uff0c\u5373\u7ba1\u9053\u673a\u5236\uff0c\u4f8b\u5982\uff1a<\/p>\n
cat xxx | awk '{xxxx}' | sort | uniq <\/code><\/pre>\n\u8fd9\u662f\u4e00\u79cd\u826f\u597d\u7684\u63a5\u53e3\u89c4\u8303\uff0c\u5de5\u5177\u7684\u529f\u80fd\u6709\u516c\u5171\u7684\u63a5\u53e3\u89c4\u8303\uff0c\u5c31\u50cf\u6d41\u6c34\u7ebf\u4e00\u6837\uff0c\u4e00\u6b65\u63a5\u7740\u4e00\u6b65\u3002\u673a\u5668\u5b66\u4e60\u7684\u5904\u7406\u8fc7\u7a0b\uff0c\u4e5f\u53ef\u4ee5\u662fpipeline\u3002\u5b9e\u9645\u4e0ascikit-learn\u5f00\u53d1\u4e86\u6574\u5957\u7684pipeline\u673a\u5236\uff0c\u5e76\u5c01\u88c5\u5230 sklearn.pipline\u547d\u540d\u7a7a\u95f4\u4e0b\u9762\u3002\u9996\u5148\uff0c\u6211\u4eec\u770b\u770b\u8fd9\u4e2a\u5e93\u90fd\u6709\u4ec0\u4e48\uff1a<\/p>\n
pipeline.FeatureUnion(transformer_list[, \u2026]) Concatenates results of multiple transformer objects.\npipeline.Pipeline(steps[, memory]) Pipeline of transforms with a final estimator.\npipeline.make_pipeline(*steps, **kwargs) Construct a Pipeline from the given estimators.\npipeline.make_union(*transformers, **kwargs) Construct a FeatureUnion from the given trans<\/code><\/pre>\n\u53ef\u4ee5\u770b\u51fa\uff0c\u6700\u5173\u952e\u7684\u662f FeatureUnion\u3001Pipeline\uff0c\u6211\u4eec\u7ee7\u7eed\u770b\u770b\u8fd92\u4e2a\u5bf9\u8c61\u90fd\u53ef\u4ee5\u5b9e\u73b0\u4ec0\u4e48\u529f\u80fd\u3002<\/p>\n
Pipeline<\/h4>\n
sklearn\u4e2d\u628a\u673a\u5668\u5b66\u4e60\u5904\u7406\u8fc7\u7a0b\u62bd\u8c61\u4e3aestimator\uff0c\u5176\u4e2destimator\u90fd\u6709fit\u65b9\u6cd5\uff0c\u8868\u793a\u201c\u5582\u201d\u6570\u636e\u8fdb\u884c\u521d\u59cb\u5316or\u8bad\u7ec3\u3002
\nestimator\u67092\u79cd\uff1a
\n1\u3001\u7279\u5f81\u53d8\u6362\uff08transformer\uff09
\n\u53ef\u4ee5\u7406\u89e3\u4e3a\u7279\u5f81\u5de5\u7a0b\uff0c\u5373\uff1a\u7279\u5f81\u6807\u51c6\u5316\u3001\u7279\u5f81\u6b63\u5219\u5316\u3001\u7279\u5f81\u79bb\u6563\u5316\u3001\u7279\u5f81\u5e73\u6ed1\u3001onehot\u7f16\u7801\u7b49
\n\u8be5\u7c7b\u578b\u7edf\u4e00\u7531\u4e00\u4e2atransform\u65b9\u6cd5\uff0c\u7528\u4e8efit\u6570\u636e\u4e4b\u540e\uff0c\u8f93\u5165\u65b0\u7684\u6570\u636e\uff0c\u8fdb\u884c\u7279\u5f81\u53d8\u6362\u3002<\/p>\n
2\u3001\u9884\u6d4b\u5668\uff08predictor\uff09
\n\u5373\u5404\u79cd\u6a21\u578b\uff0c\u6240\u6709\u6a21\u578bfit\u8fdb\u884c\u8bad\u7ec3\u4e4b\u540e\uff0c\u90fd\u8981\u7ecf\u8fc7\u6d4b\u8bd5\u96c6\u8fdb\u884cpredict\u6240\u6709\uff0c\u6709\u4e00\u4e2apredict\u7684\u516c\u5171\u65b9\u6cd5<\/p>\n
\u4e0a\u9762\u7684\u62bd\u8c61\u7684\u597d\u5904\u5373\u53ef\u5b9e\u73b0\u673a\u5668\u5b66\u4e60\u7684pipeline\uff0c\u663e\u7136\u7279\u5f81\u53d8\u6362\u662f\u53ef\u80fd\u5e76\u884c\u7684\uff08FeatureUnion\uff09\u53ef\u4ee5\u5b9e\u73b0\uff0c\u53d8\u6362\u5728\u8bad\u7ec3\u96c6\u3001\u6d4b\u8bd5\u96c6\u4e4b\u95f4\u90fd\u9700\u8981\u7edf\u4e00\uff0c\u6240\u4ee5pipeline\u53ef\u4ee5\u8fbe\u5230\u6a21\u5757\u5316\u7684\u76ee\u7684\u3002\u4e3e\u4e2aNLP\u5904\u7406\u7684\u4f8b\u5b50\uff1a<\/p>\n
# \u751f\u6210\u8bad\u7ec3\u6570\u636e\u3001\u6d4b\u8bd5\u6570\u636e\nX_train, X_test, y_train, y_test = train_test_split(X, y)\n\n# pipeline\u5b9a\u4e49\npipeline = Pipeline([\n ('vect', CountVectorizer()),\n ('tfidf', TfidfTransformer()),\n ('clf', RandomForestClassifier())\n])\n\n# train classifier\npipeline.fit(X_train, y_train)\n\n# predict on test data\ny_pred = pipeline.predict(X_test)<\/code><\/pre>\n\u663e\u7136\uff0c\u770b\u8d77\u6765pipeline\u8bad\u7ec3\u8fc7\u7a0b\u53ea\u9700\u8981fit\u548cpredict\uff0c\u5176\u5b9e\u5728pipeline\u5185\u90e8\u4f20\u8f93\u8fc7\u7a0b\uff0c\u81ea\u52a8\u8c03\u7528\u4e86fit\\transform<\/p>\n
FeatureUnion<\/h4>\n
\u4e0a\u9762\u770b\u5230\u7279\u5f81\u53d8\u6362\u5f80\u5f80\u9700\u8981\u5e76\u884c\u5316\u5904\u7406\uff0c\u5373FeatureUnion\u6240\u5b9e\u73b0\u7684\u529f\u80fd\u3002\u76f4\u63a5\u770b\u4f8b\u5b50\uff1a<\/p>\n
pipeline = Pipeline([\n('features', FeatureUnion([\n ('text_pipeline', Pipeline([\n ('vect', CountVectorizer(tokenizer=tokenize)),\n ('tfidf', TfidfTransformer())\n ])),\n ('findName', FineNameExtractor())\n])),\n\n('clf', RandomForestClassifier())\n])<\/code><\/pre>\n\u770b\u8d77\u6765\uff0cpipeline\u8fd8\u53ef\u4ee5\u5d4c\u5957pipeline\uff0c\u6574\u4e2a\u673a\u5668\u5b66\u4e60\u5904\u7406\u6d41\u7a0b\u5c31\u50cf\u6d41\u6c34\u5de5\u4eba\u4e00\u6837\u3002\u4e0a\u9762\u81ea\u5b9a\u4e49\u4e86\u4e00\u4e2apipeline\u5904\u7406\u5bf9\u8c61FineNameExtractor\uff0c\u8be5\u5bf9\u8c61\u662ftransformer\uff0c\u5b9e\u9645\u4e0a\u81ea\u5b9a\u4e49\u4e2atransformer\u662f\u5f88\u7b80\u5355\u7684\uff0c\u521b\u5efa\u4e00\u4e2a\u5bf9\u8c61\uff0c\u7ee7\u627f\u81eaBaseEstimator, TransformerMixin\u5373\u53ef\uff0c\u76f4\u63a5\u4e0a\u4ee3\u7801\uff1a<\/p>\n
from sklearn.base import BaseEstimator, TransformerMixin\nclass FineNameExtractor(BaseEstimator, TransformerMixin):\n\n def find_name(self, text):\n return True\n\n def fit(self, X, y=None):\n return self\n\n def transform(self, X):\n X_tagged = pd.Series(X).apply(self.find_name)\n return pd.DataFrame(X_tagged)<\/code><\/pre>\n\u6267\u884c\u4e00\u4e2aPIPELINE\uff0c\u53ef\u80fd\u8fd8\u5c11\u4e86\u70b9\u4ec0\u4e48\uff0c\u518d\u52a0\u4e0a\u81ea\u52a8\u8c03\u53c2\u5c31\u5b8c\u7f8e\u4e86\uff0c\u662f\u7684\uff0csklearn\u7684\u8c03\u53c2\u901a\u8fc7GridSearchCV\u5b9e\u73b0\uff0cpipeline+gridsearch\u7b80\u76f4\u662f\u7edd\u914d\u3002GridSearchCV\u5b9e\u9645\u4e0a\u4e5f\u6709fit\u3001predict\u65b9\u6cd5\uff0c\u6240\u4ee5\uff0c\u4f60\u4f1a\u53d1\u73b0\uff0c\u6574\u4e2asklearn\u7684\u673a\u5668\u5b66\u4e60\u662f\u9ad8\u6548\u62bd\u8c61\u7684\uff0c\u4ee3\u7801\u53ef\u4ee5\u5199\u7684\u5f88\u7b80\u6d01\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
pipeline\u8fd9\u4e2a\u8bcd\uff0c\u5e94\u8be5\u6765\u81ealinux\u3002\u5728linux\u4f53\u7cfb\u4e0b\u7684\u5404\u79cd\u547d\u4ee4\u5de5\u5177\u7684\u5904\u7406\uff0c\u652f\u6301pipeline\uff0c\u5373\u7ba1 […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[32,31],"class_list":["post-582","post","type-post","status-publish","format-standard","hentry","category-program","tag-pipeline","tag-31"],"_links":{"self":[{"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/posts\/582","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=582"}],"version-history":[{"count":6,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/posts\/582\/revisions"}],"predecessor-version":[{"id":613,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=\/wp\/v2\/posts\/582\/revisions\/613"}],"wp:attachment":[{"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=582"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=582"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.ahomer.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=582"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}