머신러닝 - Scikit-learn

scikit learn 순서

1. 데이터셋 불러오기
   • sklearn.datasets.load_[DATA]()
   • X, Y 가르기(Pandas 활용)
   • 열의 종류에 따라 Feature Normalization(scaling) 작업 진행

     Numeric(Continuous) Column(Variable) -> Min-Max Algorithm or Standardization 1 Categorical(Discontinuous) Column(Variable) -> one-hot encoder

2. Train/Test set
   • sklearn.model_selection.train_test_split(X,Y,test_size)

       from sklearn import model_selection
     

3. 모델객체 (Model Instance) 생성하기
   • model =
     • sklearn.linear_model.LinearRegression()
     • sklearn.linear_model.LogisticRegresion()
     • sklearn.neighbor.KNeighborsClassifier(n_neighbors)
     • sklearn.cluster.KMean(n_clusters)
     • sklearn.decomposition.PCA(n_components)
     • sklearn.svm.SVC(kernel, C, gamma)
4. 모델 학습시키기 (Model fitting)
   • model.fit(x_train, y_train)
5. 모델로 새로운 데이터 예측하기 (predict on test data)
   • model.fit(x_train, y_train)

     y = a * x + b
     coefficients == a, intercepts == b

6. 모델 테스트
   • model.predict(x_train)