machinelearning.py

import pandas as pd
import numpy as np
from sklearn.metrics import confusion_matrix, f1_score
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.preprocessing import MinMaxScaler

data = pd.read_csv("data-1711189329586.csv")
filtered_data = data[((data['target']==0)& data['mip'].between(10,100)) | ((data['target']==1)& data['mip'].between(10,100))]
print(filtered_data)
# df = pd.DataFrame(filtered_data)
# df.to_csv("Ready.csv", index = False)
# filtered_data = data[data['mip'].between(10,100)] and data
# filtered_data = filtered_data[filtered_data['target']==0]
# filtered_data = filtered_data.sort_values(by="mip")
# print("Example 1: " , filtered_data)
#
# filtered_data = data[data['mip'].between(100,300)]
# filtered_data = filtered_data[filtered_data['target']==1]
# filtered_data = filtered_data.sort_values(by="mip")
# print("Example 2: " , filtered_data)



print(filtered_data.mean(), filtered_data.max(), filtered_data.min())

filtered_data = filtered_data.sort_values(by="sip", ascending = True)
# Exersize1
 X_train, X_test, y_train, y_test = train_test_split(filtered_data.drop('target', axis=1),filtered_data['target'],test_size=0.2 ,random_state=33, stratify=filtered_data['target'], )
#
# Exersize2
scaler = MinMaxScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)
data_scaled = pd.DataFrame(X_train_scaled, columns = X_train.columns)
print(f"Максимальное значение STDC из тренировочной выборки: {X_train['stdc'].max():.3f}")
print(f"Выборочное среднее для столбца STDIP из тренировочной выборки (после нормировки): {data_scaled['stdip'].mean():.3f}")

# Exersize4(логистическая регрессия)
clf = LogisticRegression().fit(X_train_scaled, y_train)
pred = clf.predict(X_test_scaled)
print("первая матрица: ", confusion_matrix(pred, y_test))
# Exersize5 - метод ближайших соседей
knn = KNeighborsClassifier(n_neighbors=3)
knn.fit(X_train_scaled, y_train)
pred = knn.predict(X_test_scaled)
print("вторая матрица: ", confusion_matrix(pred, y_test))
print(f1_score(pred, y_test))