import sys import math import pandas as pd import seaborn as sns import numpy as np import matplotlib.pyplot as plt from scipy import stats # import plotly import plotly.express as px # import plotly.offline # import cufflinks as cf # cf.go_offline() # cf.set_config_file(offline=False, world_readable=True) # import ipywidgets as widgets # from ipywidgets import interact sns.set_style('whitegrid') %matplotlib inline %config InlineBackend.figure_format = 'retina'

import warnings warnings.filterwarnings('ignore')

print(f'Python version: {sys.version}') print(f'pandas version: {pd.__version__}') print(f'numpy version: {np.__version__}') print(f'seaborn version: {sns.__version__}') pd.Timestamp.now() # pd.Timestamp.now().strftime('%Y-%m-%d')

url = 'https://raw.githubusercontent.com/GuySuphakit/Heart-Failure-Prediction/main/heart.csv' temp_df = pd.read_csv(url) temp_df.head()

temp_df.info()

def read_data(): url = 'https://raw.githubusercontent.com/GuySuphakit/Heart-Failure-Prediction/main/heart.csv' df = pd.read_csv(url) return df # def change_lebel_of_HeartDisease(df): # # change the lebel of HeartDisease because it has caused some ambiguity. # df.HeartDisease = df.HeartDisease.replace({0: 'Normal', 1: 'Heart Disease'}) # return df def convert_FastingBS_to_category(df): df.FastingBS = df.FastingBS.astype('category') return df def convert_obj_columns_to_category(df): for c in df.columns: col_type = df[c].dtype if col_type == 'object' or col_type.name == 'category': df[c] = df[c].astype('category') return df def drop_duplicate(df): df = df.drop_duplicates() return df

df = (read_data() .pipe(convert_FastingBS_to_category) .pipe(convert_obj_columns_to_category) .pipe(drop_duplicate)) df.head()

df.info()

def print_category_columns(df): for c in df.columns: col_type = df[c].dtype if col_type.name == 'category': print(f'{c:15}: {list(enumerate(df[c].cat.categories))}') print('-' * 60) print_category_columns(df)

df.info()

def plot_mn(df, cols, n_rows:int=1, kind:str='boxplot', color='salmon'): """ plot boxplot, violin, hist in m (rows) by n (columns) >>> plot_mn(df, ['Calories', 'Fat'], 2, 'hist') """ n=len(cols) n_cols=math.ceil(n / n_rows) fig, ax = plt.subplots(n_rows, n_cols, figsize=(n_cols*3, n_rows*3.5)) ax=ax.ravel() fig.tight_layout() for i, c in enumerate(cols): col_type = df[c].dtype if col_type.name == 'category': sns.countplot(data=df, x=c, ax=ax[i]) else: if kind.lower()=='boxplot': sns.boxplot(data=df[[c]], ax=ax[i], color=color) if kind.lower()=='boxen': sns.boxenplot(data=df[[c]], ax=ax[i], color=color) elif kind.lower()=='violin': sns.violinplot(data=df[[c]], ax=ax[i], color=color) elif kind.lower()=='hist': sns.distplot(df[c], hist=True, kde=False, ax=ax[i], color=color) plot_mn(df, df.columns, 3, 'hist')

df.columns

df.describe().T.round(1)

df.describe(include='category').T

numerical= df.select_dtypes('int64').columns categorical = df.select_dtypes('category').columns print(f'Numerical Columns: {df[numerical].columns}') print('-'* 100) print(f'Categorical Columns: {df[categorical].columns}')

df.nunique().sort_values(ascending=False)

df["Age"].describe().round(1)

df.Age.mode()[0]

print(stats.skew(df.Age))

fig = px.histogram(df, x='Age', marginal='box') fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df.Sex.value_counts()

temp_df = df.groupby(by="Sex", as_index=False).agg( counts=pd.NamedAgg(column="Sex", aggfunc="count")).sort_values(by="counts", ascending=False) fig = px.bar(temp_df, x='Sex', y='counts', color='Sex', color_continuous_scale=px.colors.qualitative.D3) fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df.ChestPainType.value_counts()

temp_df = df.groupby(by="ChestPainType", as_index=False).agg( counts=pd.NamedAgg(column="ChestPainType", aggfunc="count")).sort_values(by="counts", ascending=False) fig = px.bar(temp_df, x='ChestPainType', y='counts', color='counts', color_discrete_sequence = px.colors.qualitative.D3) fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df.RestingBP.describe().round()

df.RestingBP.mode()[0]

print(stats.skew(df.RestingBP))

fig = px.histogram(df, x='RestingBP', marginal='box') fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df[df['RestingBP'] == 0]

df.Cholesterol.describe().round()

df.Cholesterol.mode()[0]

print(stats.skew(df.Cholesterol))

fig = px.histogram(df, x='Cholesterol', marginal='box') fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df[df['Cholesterol'] == 0].head()

df[df['Cholesterol'] >= 500]['Cholesterol']

df.FastingBS.value_counts()

fig = px.histogram(df, x='FastingBS', color='FastingBS') fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df["RestingECG"].value_counts()

temp_df = df.groupby(by="RestingECG", as_index=False).agg( counts=pd.NamedAgg(column="RestingECG", aggfunc="count")).sort_values(by="counts", ascending=False)

fig = px.bar(temp_df, x='RestingECG', y='counts', color='counts', color_discrete_sequence = px.colors.qualitative.D3) fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df.MaxHR.describe().round()

df.MaxHR.mode()[0]

df.MaxHR.skew()

fig = px.histogram(df, x='MaxHR', marginal='box') fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df.ExerciseAngina.value_counts()

temp_df = df.groupby(by="ExerciseAngina", as_index=False).agg( counts=pd.NamedAgg(column="ExerciseAngina", aggfunc="count")).sort_values(by="counts", ascending=False) fig = px.bar(temp_df, x='ExerciseAngina', y='counts', color='ExerciseAngina', color_continuous_scale=px.colors.qualitative.D3) fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df.Oldpeak.describe().round()

df.Oldpeak.mode()[0]

df.Oldpeak.skew()

fig = px.histogram(df, x='Oldpeak', marginal='box') fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df.ST_Slope.value_counts()

temp_df = df.groupby(by="ST_Slope", as_index=False).agg( counts=pd.NamedAgg(column="ST_Slope", aggfunc="count")).sort_values(by="counts", ascending=False) fig = px.bar(temp_df, x='ST_Slope', y='counts', color='ST_Slope', color_continuous_scale=px.colors.qualitative.D3) fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

df.HeartDisease.value_counts()

fig = px.histogram(df, x='HeartDisease', color='HeartDisease') fig.update_traces(marker_line_width=1, marker_line_color="black") fig.show()

dcorr=df[df.columns].corr() # dcorr mask = np.zeros_like(dcorr) # mask.shape mask[np.triu_indices_from(mask)] = True fig, ax = plt.subplots(figsize=(10,8)) sns.heatmap(dcorr, cmap=sns.diverging_palette(10, 145, n=100), vmin=-1, vmax=1, center=0, linewidths=1, annot=True, mask=mask, ax=ax).set_title("Correlation\nHeatmap", fontsize=22,fontweight="bold");

!pip install statsmodels

corr_value = df["Age"].corr(df["RestingBP"]) fig = px.scatter( data_frame=df, x="Age", y="RestingBP", marginal_x="histogram", marginal_y="histogram", trendline="ols", trendline_color_override="red", title=f"Correlation between Age and RestingBP is: {corr_value:.2f}", ) fig.show()

corr_value = df["Age"].corr(df["Cholesterol"]) fig = px.scatter( data_frame=df, x="Age", y="Cholesterol", marginal_x="histogram", marginal_y="histogram", trendline="ols", trendline_color_override="red", title=f"Correlation between Age and Cholester0l is: {corr_value:.2f}", ) fig.show()

corr_value = df["Age"].corr(df["Oldpeak"]) fig = px.scatter( data_frame=df, x="Age", y="Oldpeak", marginal_x="histogram", marginal_y="histogram", trendline="ols", trendline_color_override="red", title=f"Correlation between Age and Oldpeak is: {corr_value:.2f}", ) fig.show()

corr_value = df["Age"].corr(df["MaxHR"]) fig = px.scatter( data_frame=df, x="Age", y="MaxHR", marginal_x="histogram", marginal_y="histogram", trendline="ols", trendline_color_override="red", title=f"Correlation between Age and MaxHR is: {corr_value:.2f}", ) fig.show()

fig = px.density_heatmap( data_frame=df, x="Age", y="RestingBP", color_continuous_scale="PuBu" ) fig.show()

fig = px.density_heatmap( data_frame=df, x="Age", y="Cholesterol", color_continuous_scale="PuBu" ) fig.show()

fig = px.density_heatmap( data_frame=df, x="Age", y="MaxHR", color_continuous_scale="PuBu" ) fig.show()

fig = px.density_heatmap( data_frame=df, x="Age", y="Oldpeak", color_continuous_scale="PuBu" ) fig.show()

sns.pairplot(df, hue="HeartDisease", palette="husl", corner=True);

px.density_heatmap( data_frame=df, x="Age", y="RestingBP", color_continuous_scale="PuBu", facet_col="HeartDisease", title="Age vs. Chelesterol for different Heart Disease values", )

px.density_heatmap( data_frame=df, x="Age", y="Cholesterol", color_continuous_scale="PuBu", facet_col="HeartDisease", title="Age vs. Chelesterol for different Heart Disease values", )

px.density_heatmap( data_frame=df, x="Age", y="MaxHR", color_continuous_scale="PuBu", facet_col="HeartDisease", title="Age vs. Chelesterol for different Heart Disease values", )

def num_plot(df, col): fig = px.histogram(df, x=col, color="HeartDisease", marginal="box") fig.update_layout(height=400, width=700, showlegend=True) fig.update_traces(marker_line_width=1,marker_line_color="black") fig.show() cols= df.columns for col in cols: num_plot(df, col)

from sklearn.model_selection import train_test_split, KFold, StratifiedKFold, ShuffleSplit, cross_val_score, cross_val_predict, cross_validate from sklearn.metrics import confusion_matrix, plot_confusion_matrix, classification_report, accuracy_score, precision_score, recall_score, precision_recall_fscore_support, f1_score

feature_cols = ['Age', 'Sex', 'ChestPainType', 'RestingBP', 'Cholesterol', 'FastingBS', 'RestingECG', 'MaxHR', 'ExerciseAngina', 'Oldpeak', 'ST_Slope'] target_cols = 'HeartDisease' X = df[feature_cols] y = df[target_cols] X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.15, random_state=42, stratify=y)

def data_preprocessing(X_train): ''' replace outliers with mean to make the distribution more normal. ''' mean_chol = X_train[X_train.Cholesterol > 0]['Cholesterol'].mean() mean_rest = X_train[X_train.RestingBP > 0]['Cholesterol'].mean() X_train['Cholesterol'].replace(to_replace = [0,X_train[X_train['Cholesterol'] >= 500]['Cholesterol']], value = mean_chol, inplace = True) X_train['RestingBP'].replace(to_replace = [0,244.635389], value = mean_rest, inplace = True) return X_train X_train = data_preprocessing(X_train)

df.columns

y_train.value_counts()

!pip install flaml

from flaml import AutoML automl = AutoML()

settings = { "time_budget": 180, # total running time in seconds "metric": 'accuracy', # can be: 'r2', 'rmse', 'mae', 'mse', 'accuracy', 'roc_auc', 'roc_auc_ovr', # 'roc_auc_ovo', 'log_loss', 'mape', 'f1', 'ap', 'ndcg', 'micro_f1', 'macro_f1' "task": 'classification', # task type "log_file_name": 'airlines_experiment.log', # flaml log file }

automl.fit(X_train=X_train, y_train=y_train, **settings)

''' retrieve best config and best learner''' print('Best ML leaner:', automl.best_estimator) print('Best hyperparmeter config:', automl.best_config) print('Best accuracy on validation data: {0:.4g}'.format(1-automl.best_loss)) print('Training duration of best run: {0:.4g} s'.format(automl.best_config_train_time))

automl.model

''' compute predictions of testing dataset ''' y_pred = automl.model.predict(X_test) print('Predicted labels', y_pred) print('True labels', y_test) y_pred_proba = automl.model.predict_proba(X_test)[:,1]

''' compute different metric values on testing dataset''' from flaml.ml import sklearn_metric_loss_score print('accuracy', '=', 1 - sklearn_metric_loss_score('accuracy', y_pred, y_test)) print('roc_auc', '=', 1 - sklearn_metric_loss_score('roc_auc', y_pred_proba, y_test)) print('log_loss', '=', sklearn_metric_loss_score('log_loss', y_pred_proba, y_test)) print('Training duration of best run: {0:.4g} s'.format(automl.best_config_train_time))

cm = confusion_matrix(y_test, automl.model.predict(X_test)) cm

plot_confusion_matrix(automl.model, X_test, y_test, cmap='Oranges', normalize='true')

import shap print(f'shap version {shap.__version__}') # load JS visualization code to notebook shap.initjs()

def case_detail(case_data): ''' format obj returned from shap.force_plot() ''' de=pd.DataFrame(case_data.data['features']) fcols=[] for i in case_data.data['features'].keys(): fcols.append(case_data.data['featureNames'][i]) de.columns=fcols return de def individual_case_plot(explainer, X, case_index, verbose=False): """ >>> individual_case_plot(explainer, X_train, 1) """ shap_values = explainer.shap_values(X.iloc[[case_index]]) g=shap.force_plot(explainer.expected_value, shap_values=shap_values, features=X.iloc[case_index, :]) if verbose: pprint(g.__dict__) return g

# explain the model's predictions using SHAP explainer = shap.TreeExplainer(automl.model) shap_values = explainer.shap_values(X_test)

X

shap_values[:3]

shap.summary_plot(shap_values, X_train, plot_type="bar")

feature_cols

dshap=pd.DataFrame(shap_values, columns=feature_cols) dshap