Machine Learning & Deep Learning practice code.1

• Installing and calling Seaborn package

!pip install seaborn
import seaborn as sns
import matplotlib.pyplot as plt

• Reading the csv file

df = pd.read_csv("your file name", encoding='cp949')

• Checking Correlation between of the data using Heatmap

sns.heatmap(df.corr(), annot=True)

• Setting the font

plt.rc('font', family='Malgun Gothic')
plt.show()

• Get the average height for each age group in bar code

df.columns

sns.barplot(x='gentercode', y='drinkingornot', data=df)

plt.rc('font', family='Malgun Gothic')  # assigning the font
plt.show()

• Handling of Missing Values
• Deleting dental related data
  ( 27...
  28...
  29...
  30...)

df.head(1)

pd.set_option('display.max_columns', None)  # printng every columns
pd.set_option('display.max_row', None) # printing every rows

df.head(1)

df.columns

df = df.drop(['column names stated above1', '2', '3', '4'], axis=1)

df['baseyear'].nunique()

df['subscriber serial number'].nunique()

df = df.drop(['base year', 'subsciber serial number'], axis=1)

df.head()

df = df.fillna(0) # need to write another way of dealing this 

• Convert gender code and age code data with One Hot Incoding

df_dummies=pd.get_dummies(df[['gender code', 'trial code']], drop_first=True)

df_dummies.head(1)

df.head(1)

df1 = pd.get_dummies(data=df, columns=cal_cols)

df.info()

df = pd.get_dummies(df, columns=['trail code', 'gender code'])

df.head(1)

• Adding whether drunk or not. Set the remaining columns to Featrue
  and Divide by 7:3

y = df.drunk

X = df.drop(['drunk'], axis=1)

from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, scratify = y, random_state=42)

• Apply scaling to the data (Standard Scaler)

from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()

scaler.fit(X_train)

X_train = scaler.transform(X_train)

X_test = scaler.transform(X_test)

• Train with RandomForest model

from sklearn.ensemble import RandomForestClassifier

rfc = RandomForestClassifier(n_estimators=30, max_features=3, max_depth=7, random_state=21)
rfc.fit(X_train, y_train)

rfc.score(X_train, y_train)

• Evaluating the performance (Get the confusion matrix and visualize it with a heatmap graph)

from sklearn.metrics import confusion_matrix, classification_report

y_pred = rfc.predict(X_test)
conf_mat = confusion_matrix(y_test, y_pred)

sns.heatmap(conf_mat, annot=True)

plt.show()

print(f"classification_report: {classification_report(y_test, y_pred)}")

• Making a deep-learning model (More than X hidden layers.
  Dropout configuration required.
  Apply Early Stopping Model Checkpoint)

import tensorflow as tf

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Activation, Dropout
from tensorflow.keras.callbacks import ModelCheckpoint, EarlyStopping

X_train.shape

model = Sequential()
model.add(Dense(128, activation = 'relu', input_shape=(X_train.shape[1],)))
model.add(Dropout(0.5))
model.add(Dense(128, activation = 'relu'))
model.add(Dropout(0.5))
model.add(Dense(64, activation = 'relu'))
model.add(Dropout(0.5))
model.add(Dense(32, activation = 'relu'))
model.add(Dropout(0.5))
model.add(Dense(1, activation='sigmoid'))

model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['acc'])
mc = ModelCheckpoint(filepath='best.h5', verbose=1, monitor='val_loss', mode='auto', save_best_only=True)
es = EarlyStopping(monitor='val_loss', mode='auto', verbose=1, patience=5)

history= model.fit(X_train,y_train, epochs=50,
                   validation_data=(X_test,y_test),
                #    validation_split=0.1,
                    verbose=1,
                    callbacks=[es,mc]
                   )

print(history['acc'])

print(history.history['acc'])

plt.figure(figsize=(10,5))
plt.plot(history.history['acc'], 'red', label="acc")
plt.plot(history.history['val_acc'], 'blue', label='val_acc')

plt.title("타이틀이에요")
plt.xlabel('학습회수')
plt.ylabel('정확도')
plt.rc('font', family='Malgun Gothic') #setting the font
plt.show()