LDA

Implemented LDA introduced at “Introduction to Statistical Learning: With Applications in R. Chapter 4 Lab”

After we got the result from logistric regression, LDA is tried to get the better result.

Using scikit-learn

Direction as a response, "Lag1" and "Lag2" as predictors.

from sklearn.lda import LDA
lda = LDA()
lda.fit(X_train, y_resp_train)
lda.score(X_test, y_resp_test)
# 0.55952380952380953 correct rate

QDA

Using scikit-learn

from sklearn.qda import QDA
qda = QDA()
qda.fity(X_train, _resp_train)
qda.score(X_test, y_resp_test)
# 0.59920634920634919 correct rate

qda.means_
qda.priors_
qda.rotations_ 
qda.classes_ 
lda. priors_ 
lda. coef_ 

KNN

Using scikit-learn

from sklearn.neighbors import KNeighborsClassifier
knn = KNeighborsClassifier(3)
knn.fit(X_train, y_resp_train)
knn.predict(X_test)
knn.score(X_test,y_resp_test)
# 0.53174603174603174 correct rate

From this dataset and classification methods, QDA has the best correct rate.

[References] [1] James, Gareth, Daniela Witten, Trevor Hastie, and Robert Tibshirani. An Introduction to Statistical Learning: With Applications in R. Print.

[2] Hauck, Trent. Scikit-learn Cookbook. Packt, 2014. Print.