https://mediawiki.zeropage.org/api.php?action=feedcontributions&feedformat=atom&user=211.104.144.68 2026-05-15T06:00:08Z User contributions MediaWiki 1.39.8 https://mediawiki.zeropage.org/index.php?title=%EB%A8%B8%EC%8B%A0%EB%9F%AC%EB%8B%9D%EC%8A%A4%ED%84%B0%EB%94%94/2016/2016_07_23&diff=50332 2016-07-26T11:15:30Z

<p>211.104.144.68: </p> <hr /> <div>[[머신러닝스터디/2016]]<br /> [[머신러닝스터디/2016/목차]]<br /> == 내용 ==<br /> * SVM 실습 with [http://scikit-learn.org/stable/ sklearn]<br /> ** skflow(tensorflow의 contib/learn으로 흡수됨)을 사용하려 했으나 svm모듈 부분이 최신 커밋에만 포함되어 있어 sklearn을 사용하기로 함<br /> ** sklearn 버전은 0.17.1<br /> 설치 방법<br /> $ sudo pip install sklearn<br /> <br /> === 코드 ===<br /> import sklearn<br /> from sklearn import svm<br /> <br /> #### SVC with rbf kernel<br /> # default is rbf kernel<br /> clf = svm.SVC()<br /> x_data = [[0,0], [0,1], [1,0], [1,1]]<br /> <br /> # linear<br /> ## And gate<br /> y_data = [0, 0, 0, 1]<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0]) # wrong<br /> <br /> ## Or gate<br /> y_data = [0, 1, 1, 1]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([1, 1, 1, 1])<br /> <br /> # non-linear<br /> ## Xor gate<br /> y_data = [0, 1, 1, 0]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([0, 1, 1, 0]) # Correct answer<br /> <br /> <br /> #### SVC with Linear kernel<br /> clf = svm.SVC(kernel='linear')<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='linear',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0])<br /> <br /> <br /> #### LinearSVC<br /> clf = svm.LinearSVC()<br /> clf.fit(x_data, y_data)<br /> # LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,<br /> # intercept_scaling=1, loss='squared_hinge', max_iter=1000,<br /> # multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,<br /> # verbose=0)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 1]) # Correct answer<br /> <br /> * non-linear는 분류하지만 linear는 분류하지 못하는 SVC?<br /> * multi-class에서 분류 시에 voter의 결과에 따라 분류됨<br /> ** 결과에 0이 많으면(And gate) 모두 0이 되고 <br /> y_data = [0, 0, 0, 1]면 결과가 [0, 0, 0, 0]<br /> ** 1이 더 많은 경우(Or gate) 학습 결과가 다음과 같음<br /> y_data = [0, 1, 1, 1]면 결과가 [1, 1, 1, 1]<br /> * SVC와 LinearSVC의 차이란?<br /> * [http://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html#sklearn-svm-svc]에 따르면 둘의 차이점은 다음과 같다<br /> * SVC; C-Support Vector Classification. The implementation is based on libsvm. <br /> * LinearSVC; Scalable Linear Support Vector Machine for classification implemented using liblinear. Check the See also section of LinearSVC for more comparison element.<br /> <br /> == 다음 시간에는 ==<br /> == 더 보기 ==<br /> </div>

211.104.144.68 https://mediawiki.zeropage.org/index.php?title=%EB%A8%B8%EC%8B%A0%EB%9F%AC%EB%8B%9D%EC%8A%A4%ED%84%B0%EB%94%94/2016/2016_07_23&diff=50331 2016-07-26T10:45:57Z

<p>211.104.144.68: </p> <hr /> <div>[[머신러닝스터디/2016]]<br /> [[머신러닝스터디/2016/목차]]<br /> == 내용 ==<br /> * SVM 실습 with [http://scikit-learn.org/stable/ sklearn]<br /> ** skflow(tensorflow의 contib/learn으로 흡수됨)을 사용하려 했으나 svm모듈 부분이 최신 커밋에만 포함되어 있어 sklearn을 사용하기로 함<br /> ** sklearn 버전은 0.17.1<br /> 설치 방법<br /> $ sudo pip install sklearn<br /> <br /> === 코드 ===<br /> import sklearn<br /> from sklearn import svm<br /> <br /> #### SVC with rbf kernel<br /> # default is rbf kernel<br /> clf = svm.SVC()<br /> x_data = [[0,0], [0,1], [1,0], [1,1]]<br /> <br /> # linear<br /> ## And gate<br /> y_data = [0, 0, 0, 1]<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0]) # wrong<br /> <br /> ## Or gate<br /> y_data = [0, 1, 1, 1]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([1, 1, 1, 1])<br /> <br /> # non-linear<br /> ## Xor gate<br /> y_data = [0, 1, 1, 0]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([0, 1, 1, 0]) # Correct answer<br /> <br /> <br /> #### SVC with Linear kernel<br /> clf = svm.SVC(kernel='linear')<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='linear',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0])<br /> <br /> <br /> #### LinearSVC<br /> clf = svm.LinearSVC()<br /> clf.fit(x_data, y_data)<br /> # LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,<br /> # intercept_scaling=1, loss='squared_hinge', max_iter=1000,<br /> # multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,<br /> # verbose=0)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 1]) # Correct answer<br /> <br /> * non-linear는 분류하지만 linear는 분류하지 못하는 SVC?<br /> * multi-class에서 분류 시에 voter의 결과에 따라 분류됨<br /> ** 결과에 0이 많으면(And gate) 모두 0이 되고 <br /> y_data = [0, 0, 0, 1]면 결과가 [0, 0, 0, 0]<br /> ** 1이 더 많은 경우(Or gate) 학습 결과가 다음과 같음<br /> y_data = [0, 1, 1, 1]면 결과가 [1, 1, 1, 1]<br /> <br /> == 다음 시간에는 ==<br /> == 더 보기 ==<br /> </div>

211.104.144.68 https://mediawiki.zeropage.org/index.php?title=%EB%A8%B8%EC%8B%A0%EB%9F%AC%EB%8B%9D%EC%8A%A4%ED%84%B0%EB%94%94/2016/2016_07_23&diff=50330 2016-07-26T10:44:58Z

<p>211.104.144.68: </p> <hr /> <div>[[머신러닝스터디/2016]]<br /> [[머신러닝스터디/2016/목차]]<br /> == 내용 ==<br /> * SVM 실습 with [http://scikit-learn.org/stable/ sklearn]<br /> ** skflow(tensorflow의 contib/learn으로 흡수됨)을 사용하려 했으나 svm모듈 부분이 최신 커밋에만 포함되어 있어 sklearn을 사용하기로 함<br /> ** sklearn 버전은 0.17.1<br /> 설치 방법<br /> $ sudo pip install sklearn<br /> <br /> === 코드 ===<br /> import sklearn<br /> from sklearn import svm<br /> <br /> #### SVC with rbf kernel<br /> # default is rbf kernel<br /> clf = svm.SVC()<br /> x_data = [[0,0], [0,1], [1,0], [1,1]]<br /> <br /> # linear<br /> ## And gate<br /> y_data = [0, 0, 0, 1]<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0]) # wrong<br /> <br /> ## Or gate<br /> y_data = [0, 1, 1, 1]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([1, 1, 1, 1])<br /> <br /> # non-linear<br /> ## Xor gate<br /> y_data = [0, 1, 1, 0]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([0, 1, 1, 0]) # Correct answer<br /> <br /> <br /> #### SVC with Linear kernel<br /> clf = svm.SVC(kernel='linear')<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='linear',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0])<br /> <br /> <br /> #### LinearSVC<br /> clf = svm.LinearSVC()<br /> clf.fit(x_data, y_data)<br /> # LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,<br /> # intercept_scaling=1, loss='squared_hinge', max_iter=1000,<br /> # multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,<br /> # verbose=0)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 1]) # Correct answer<br /> <br /> * non-linear는 분류하지만 linear는 분류하지 못하는 SVC?<br /> * multi-class에서 분류 시에 voter의 결과에 따라 분류됨<br /> ** 결과에 0이 많으면(And gate) 모두 0이 되고 <br /> y_data = [0, 0, 0, 1]면 결과가 [0, 0, 0, 0]<br /> ** 1이 더 많으면(Or gate) <br /> y_data = [0, 1, 1, 1]면 결과가 [1, 1, 1, 1]<br /> * and gate는multi class가 아님<br /> == 다음 시간에는 ==<br /> == 더 보기 ==<br /> </div>

211.104.144.68 https://mediawiki.zeropage.org/index.php?title=%EB%A8%B8%EC%8B%A0%EB%9F%AC%EB%8B%9D%EC%8A%A4%ED%84%B0%EB%94%94/2016/2016_07_23&diff=50329 2016-07-26T10:17:57Z

<p>211.104.144.68: </p> <hr /> <div>[[머신러닝스터디/2016]]<br /> [[머신러닝스터디/2016/목차]]<br /> == 내용 ==<br /> * SVM 실습 with [http://scikit-learn.org/stable/ sklearn]<br /> ** skflow(tensorflow의 contib/learn으로 흡수됨)을 사용하려 했으나 svm모듈 부분이 최신 커밋에만 포함되어 있어 sklearn을 사용하기로 함<br /> ** sklearn 버전은 0.17.1<br /> 설치 방법<br /> $ sudo pip install sklearn<br /> <br /> === 코드 ===<br /> import sklearn<br /> from sklearn import svm<br /> <br /> #### SVC with rbf kernel<br /> # default is rbf kernel<br /> clf = svm.SVC()<br /> x_data = [[0,0], [0,1], [1,0], [1,1]]<br /> <br /> # linear<br /> y_data = [0, 0, 0, 1]<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0]) # wrong<br /> <br /> # non-linear<br /> y_data = [0, 1, 1, 0]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([0, 1, 1, 0]) # Correct answer<br /> <br /> <br /> #### SVC with Linear kernel<br /> clf = svm.SVC(kernel='linear')<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='linear',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0])<br /> <br /> <br /> #### LinearSVC<br /> clf = svm.LinearSVC()<br /> clf.fit(x_data, y_data)<br /> # LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,<br /> # intercept_scaling=1, loss='squared_hinge', max_iter=1000,<br /> # multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,<br /> # verbose=0)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 1]) # Correct answer<br /> == 다음 시간에는 ==<br /> == 더 보기 ==<br /> </div>

211.104.144.68 https://mediawiki.zeropage.org/index.php?title=%EB%A8%B8%EC%8B%A0%EB%9F%AC%EB%8B%9D%EC%8A%A4%ED%84%B0%EB%94%94/2016/2016_07_23&diff=50328 2016-07-26T10:17:47Z

<p>211.104.144.68: </p> <hr /> <div>[[머신러닝스터디/2016]]<br /> [[머신러닝스터디/2016/목차]]<br /> == 내용 ==<br /> * SVM 실습 with [http://scikit-learn.org/stable/ sklearn]<br /> ** skflow(tensorflow의 contib/learn으로 흡수됨)을 사용하려 했으나 svm모듈 부분이 최신 커밋에만 포함되어 있어 sklearn을 사용하기로 함<br /> ** sklearn 버전은 0.17.1<br /> * 설치 방법<br /> $ sudo pip install sklearn<br /> <br /> === 코드 ===<br /> import sklearn<br /> from sklearn import svm<br /> <br /> #### SVC with rbf kernel<br /> # default is rbf kernel<br /> clf = svm.SVC()<br /> x_data = [[0,0], [0,1], [1,0], [1,1]]<br /> <br /> # linear<br /> y_data = [0, 0, 0, 1]<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0]) # wrong<br /> <br /> # non-linear<br /> y_data = [0, 1, 1, 0]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([0, 1, 1, 0]) # Correct answer<br /> <br /> <br /> #### SVC with Linear kernel<br /> clf = svm.SVC(kernel='linear')<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='linear',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0])<br /> <br /> <br /> #### LinearSVC<br /> clf = svm.LinearSVC()<br /> clf.fit(x_data, y_data)<br /> # LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,<br /> # intercept_scaling=1, loss='squared_hinge', max_iter=1000,<br /> # multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,<br /> # verbose=0)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 1]) # Correct answer<br /> == 다음 시간에는 ==<br /> == 더 보기 ==<br /> </div>

211.104.144.68 https://mediawiki.zeropage.org/index.php?title=%EB%A8%B8%EC%8B%A0%EB%9F%AC%EB%8B%9D%EC%8A%A4%ED%84%B0%EB%94%94/2016/2016_07_23&diff=50327 2016-07-26T10:17:28Z

<p>211.104.144.68: </p> <hr /> <div>[[머신러닝스터디/2016]]<br /> [[머신러닝스터디/2016/목차]]<br /> == 내용 ==<br /> * SVM 실습 with [http://scikit-learn.org/stable/ sklearn]<br /> ** skflow(tensorflow의 contib/learn으로 흡수됨)을 사용하려 했으나 svm모듈 부분이 최신 커밋에만 포함되어 있어 sklearn을 사용하기로 함<br /> ** sklearn 버전은 0.17.1<br /> $ sudo pip install sklearn<br /> <br /> <br /> === 코드 ===<br /> import sklearn<br /> from sklearn import svm<br /> <br /> #### SVC with rbf kernel<br /> # default is rbf kernel<br /> clf = svm.SVC()<br /> x_data = [[0,0], [0,1], [1,0], [1,1]]<br /> <br /> # linear<br /> y_data = [0, 0, 0, 1]<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0]) # wrong<br /> <br /> # non-linear<br /> y_data = [0, 1, 1, 0]<br /> clf.fit(x_data, y_data)<br /> clf.predict(x_data)<br /> # array([0, 1, 1, 0]) # Correct answer<br /> <br /> <br /> #### SVC with Linear kernel<br /> clf = svm.SVC(kernel='linear')<br /> clf.fit(x_data, y_data)<br /> # SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,<br /> # decision_function_shape=None, degree=3, gamma='auto', kernel='linear',<br /> # max_iter=-1, probability=False, random_state=None, shrinking=True,<br /> # tol=0.001, verbose=False)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 0])<br /> <br /> <br /> #### LinearSVC<br /> clf = svm.LinearSVC()<br /> clf.fit(x_data, y_data)<br /> # LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,<br /> # intercept_scaling=1, loss='squared_hinge', max_iter=1000,<br /> # multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,<br /> # verbose=0)<br /> clf.predict(x_data)<br /> # array([0, 0, 0, 1]) # Correct answer<br /> == 다음 시간에는 ==<br /> == 더 보기 ==<br /> </div>

211.104.144.68