hershyz / classificationlib Goto Github PK

# importing all library modules:
import dataframe
import correlation_module
import sqrt_distance_classifier
import abs_distance_classifier
import percent_distance_classifier
import stddev_distance_classifier
import knn
import classnet

df = dataframe.Dataframe('test-data/drug200.csv')
correlation_module.run(df)

 # (dataframe, input features)
model = sqrt_distance_classifier.train(df, ['Na_to_K', 'Cholesterol', 'BP'])
 # (model, dataframe, output feature)
print(sqrt_distance_classifier.eval(model, df, 'Drug')) # get accuracy

output: 0.7

point = {
    'Age': 23,
    'Sex': 'F',
    'BP': 'HIGH',
    'Cholesterol': 'HIGH',
    'Na_to_K': 25.355
}
point_numerical = df.convert_point(point)
# (converted numerical point, model)
print(sqrt_distance_classifier.predict(point_numerical, model))

output: DrugY

# (dataframe, input features)
model = abs_distance_classifier.train(df, ['Na_to_K', 'BP', 'Sex'])
# (model, dataframe, output feature)
print(abs_distance_classifier.eval(model, df, 'Drug')) # get accuracy

output: 0.7

point = {
    'Age': 23,
    'Sex': 'F',
    'BP': 'HIGH',
    'Cholesterol': 'HIGH',
    'Na_to_K': 25.355
}
point_numerical = df.convert_point(point)
# (converted numerical point, model)
print(abs_distance_classifier.predict(point_numerical, model))

output: DrugY

# (dataframe, input features)
model = percent_distance_classifier.train(df, ['Na_to_K', 'BP', 'Age'])
# (model, dataframe, output feature)
print(percent_distance_classifier.eval(model, df, 'Drug')) # get accuracy

output: 0.715

point = {
    'Age': 23,
    'Sex': 'F',
    'BP': 'HIGH',
    'Cholesterol': 'HIGH',
    'Na_to_K': 25.355
}
point_numerical = df.convert_point(point)
# (converted numerical point, model)
print(percent_distance_classifier.predict(point_numerical, model))

output: DrugY

# (dataframe, input features)
model = stddev_distance_classifier.train(df, ['Na_to_K', 'BP', 'Age'])
# (model, dataframe, output feature)
print(stddev_distance_classifier.eval(model, df, 'Drug')) # get accuracy

output: 0.705

point = {
    'Age': 23,
    'Sex': 'F',
    'BP': 'HIGH',
    'Cholesterol': 'HIGH',
    'Na_to_K': 25.355
}
point_numerical = df.convert_point(point)
# (converted numerical point, model)
print(stddev_distance_classifier.predict(point_numerical, model))

output: DrugY

# (dataframe, input features)
model = classnet.train(df, ['Na_to_K', 'Cholesterol', 'Sex', 'BP', 'Age'])
# (model, dataframe, output feature)
print(classnet.eval(model, df, 'Drug')) # get accuracy

output: 0.24

point = {
    'Age': 23,
    'Sex': 'F',
    'BP': 'HIGH',
    'Cholesterol': 'HIGH',
    'Na_to_K': 25.355
}
point_numerical = df.convert_point(point)
# (converted numerical point, model)
print(classnet.predict(point_numerical, model))

output: DrugC

df = dataframe.Dataframe('test-data/drug200.csv')
# (dataframe, input features, training dataset ratio)
model = knn.train(df, ['Na_to_K', 'Cholesterol', 'BP'], 'Drug', 0.5)
# (model, dataframe, output feature)
print(knn.eval(model, df, 'Drug'))

output: 0.88

point = {
    'Age': 23,
    'Sex': 'F',
    'BP': 'HIGH',
    'Cholesterol': 'HIGH',
    'Na_to_K': 25.355
}
point_numerical = df.convert_point(point)
# (converted numerical point, output feature, model)
print(knn.predict(point_numerical, 'Drug', model))

output: DrugY