model.py

# -*- coding: utf-8 -*-
"""BlockchainProject.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1OlpfSxD34eNd3R-Ft9e9iV-9jnCF6Pmm
"""

import pandas as pd
import imblearn

print(imblearn.__version__)
from sklearn.preprocessing import StandardScaler
from sklearn.svm import SVC
from sklearn.model_selection import train_test_split
from imblearn.over_sampling import SMOTE

num_features = [
    'hemoglobin',
    'platelets',
    'MPV',
    'RBC',
    'lymphocytes',
    'MCHC',
    'leukocytes',
    'basophils',
    'MCH',
    'eosinophils',
    'MCV',
    'monocytes',
    'RDW']
scaler = StandardScaler()


def algo(df, dp):
    df = pd.DataFrame(df)
    df[num_features] = scaler.fit_transform(df[num_features])

    # oversample = RandomOverSampler(sampling_strategy=1)

    smt = SMOTE(sampling_strategy=0.5, k_neighbors=5, random_state=1206)

    y = df['covid19_Res']
    X = df.drop(['covid19_Res'], axis=1)
    X.head()

    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=1206,
                                                        stratify=y)

    X_over, y_over = smt.fit_resample(X_train, y_train)

    hyperPara = {'C': [10], 'gamma': [0.1], 'kernel': ['rbf'], 'shrinking': [True]}
    # svm1 = SVC(C=10, gamma=0.1, kernel='rbf')

    svm1 = SVC(C=10, gamma=0.1, kernel='rbf', class_weight={1: 10})

    svm1.fit(X_train, y_train)

    y_pred = svm1.predict(X_test)

    y_pred = svm1.predict(dp)

    print(f'You are\n {dp}')
    if (y_pred[0] == 0):
        y_pred = "Negative"
    elif (y_pred[0] == 1):
        y_pred = "Positive"
    print(f'Current records number is {df.shape[0]}')
    print(f'The AI predicts you are {y_pred}!')
    from sklearn.metrics import classification_report

    y_pred_2 = svm1.predict(X_test)
    # print(f'Result is {classification_report(y_test, y_pred_2)}')
    return y_pred