classification_detector

a program for find the best algorithm between classification algorithms(KNN, Decision tree, logistic regression, SVM)

in this program you can give your data and see which of this algorithms have better score.

this is an example: classification(data=df, features=['age', 'sex','trtbps', 'chol','cp','fbs','restecg', 'thalachh','oldpeak', 'slp','exng', 'caa','thall'], y_ouput='output', wanna_normalize=False,goal='negFal', number_of_repeats=5)

classification: this is the main function.

data: this gives the data as dataFrame.

features: this gives the list of Xs.

y_ouput: this gives the name of y column. Note: the values of this columns should be in binary(0, 1) format.

goal: this is the score type we want to evaluate the model by it. the goal can be these: F1-score: the highest f1-score.(default) negFal: the lowest negative False percent. posFal: the lowest positive False percent. negtru: the highest negative True percent. postru: the highest positive True percent. jac : jaccard-score.

wanna_normalize: this show we want to normalize our data or not(default:True).

number_of_repeats: this show the number of test we want to do. for example number_of_repeats=3 indicates that first we randomize Xs and calculate the score of algorithms then again randomize the Xs and calculate scores and so on for 3 times. this is for average score.(default=1)

output of the function: the best F1-score score for 1 states is: 0.802214, for state 1, with algorithm KNN with k: 9.

the best average of F1-score scores for 1 states is: 0.802214 with algorithm KNN.