## <i> Project 2: Supervised Machine Learning </i>
### **_by Sebastian Sbirna_**
***
In this report, we will evaluate the performance and characteristics of various types of supervised learning models upon our chosen Heart Disease dataset. For more information upon the actual data dictionary and a description of the properties of our observations and attributes, please refer to our former project [1].
### I. Regression Models
```python
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.pyplot import figure, plot, xlabel, ylabel, legend, show, clim, semilogx, loglog, title, subplot, grid
import sklearn.linear_model as lm
from sklearn.linear_model import Ridge, LinearRegression, LogisticRegression
from sklearn import model_selection, tree
from scipy import stats
import torch
from toolbox_02450 import feature_selector_lr, bmplot, rlr_validate, train_neural_net
import warnings
warnings.filterwarnings('ignore')
%matplotlib inline
```
### Regression, part A
Our dataset was collected for classification purposes, having 14 attributes: 5 numerical and 9 categorical. Out of the 14 attributes, there exists one ‘target’ variable which should be predicted in a classification setting. Otherwise, there are no variables which were collected specifically for a regression purpose within this dataset, therefore model results within this section may be prone to more errors.
Still, we will use one of the five numerical (_of ratio type_) attributes as our criterion (_i.e. dependent_) variable for a regression analysis, where the other 13 attributes (_which increase to 20 attributes after a one-out-of-K encoding_) will all act as predictor (_i.e. independent_) variables. Now, we must decide which variable is best suited to serve as criterion.
However, our dataset has one important peculiarity which makes regression extremely difficult: most of our attributes are either uncorrelated or weakly correlated with each other, resulting in very little predictive power being stored in our dataset for any variable.
#### Loading the dataset and performing data wrangling:
```python
df = pd.read_csv('heart.csv')
```
```python
df.head()
```
<div>
<style scoped>
.dataframe tbody tr th:only-of-type {
vertical-align: middle;
}
.dataframe tbody tr th {
vertical-align: top;
}
.dataframe thead th {
text-align: right;
}
</style>
<table border="1" class="dataframe">
<thead>
<tr style="text-align: right;">
<th></th>
<th>age</th>
<th>sex</th>
<th>cp</th>
<th>trestbps</th>
<th>chol</th>
<th>fbs</th>
<th>restecg</th>
<th>thalach</th>
<th>exang</th>
<th>oldpeak</th>
<th>slope</th>
<th>ca</th>
<th>thal</th>
<th>target</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>63</td>
<td>1</td>
<td>3</td>
<td>145</td>
<td>233</td>
<td>1</td>
<td>0</td>
<td>150</td>
<td>0</td>
<td>2.3</td>
<td>0</td>
<td>0</td>
<td>1</td>
<td>1</td>
</tr>
<tr>
<th>1</th>
<td>37</td>
<td>1</td>
<td>2</td>
<td>130</td>
<td>250</td>
<td>0</td>
<td>1</td>
<td>187</td>
<td>0</td>
<td>3.5</td>
<td>0</td>
<td>0</td>
<td>2</td>
<td>1</td>
</tr>
<tr>
<th>2</th>
<td>41</td>
<td>0</td>
<td>1</td>
<td>130</td>
<td>204</td>
<td>0</td>
<td>0</td>
<td>172</td>
<td>0</td>
<td>1.4</td>
<td>2</td>
<td>0</td>
<td>2</td>
<td>1</td>
</tr>
<tr>
<th>3</th>
<td>56</td>
<td>1</td>
<td>1</td>
<td>120</td>
<td>236</td>
<td>0</td>
<td>1</td>
<td>178</td>
<td>0</td>
<td>0.8</td>
<td>2</td>
<td>0</td>
<td>2</td>
<td>1</td>
</tr>
<tr>
<th>4</th>
<td>57</td>
<td>0</td>
<td>0</td>
<td>120</td>
<td>354</td>
<td>0</td>
<td>1</td>
<td>163</td>
<td>1</td>
<td>0.6</td>
<td>2</td>
<td>0</td>
<td>2</td>
<td>1</td>
</tr>
</tbody>
</table>
</div>
```python
df.drop(index = (df[df.ca == 4]).index, inplace = True)
df.drop(index = (df[df.thal == 0]).index, inplace = True)
df.loc[df.thal == 1, 'thal'] = 6
df.loc[df.thal == 3, 'thal'] = 7
df.loc[df.thal == 2, 'thal'] = 3
df.loc[df.cp == 0, 'cp'] = 4
df.loc[df.cp == 3, 'cp'] = 7
df.loc[df.cp == 2, 'cp'] = 3
df.loc[df.cp == 1, 'cp'] = 2
df.loc[df.cp == 7, 'cp'] = 1
df.loc[df.slope == 2, 'slope'] = 3
df.loc[df.slope == 1, 'slope'] = 2
df.loc[df.slope == 0, 'slope'] = 1
```
```python
numerical_columns = ['age', 'trestbps', 'chol', 'thalach', 'oldpeak']
```
```python
df['sex_male'] = df.sex
df.drop(columns = 'sex', inplace = True)
df = pd.get_dummies(data = df, columns = ['cp'], drop_first=True)
df.rename({'cp_2': 'cp_atypical', 'cp_3' : 'cp_non_anginal', 'cp_4': 'cp_asymptomatic'}, axis = 'columns', inplace = True)
df['fbs_true'] = df.fbs
df.drop(columns = 'fbs', inplace = True)
df = pd.get_dummies(data = df, columns = ['restecg'], drop_first=True)
df.rename({'restecg_1': 'restecg_st_t', 'restecg_2' : 'restecg_hypertrophy'}, axis = 'columns', inplace = True)
df['exang_yes'] = df.exang
df.drop(columns = 'exang', inplace = True)
df = pd.get_dummies(data = df, columns = ['slope'], drop_first=True)
df.rename({'slope_2': 'slope_flat', 'slope_3' : 'slope_downsloping'}, axis = 'columns', inplace = True)
df = pd.get_dummies(data = df, columns = ['ca'], drop_first=True)
df = pd.get_dummies(data = df, columns = ['thal'], drop_first=True)
df.rename({'thal_6': 'thal_fixed', 'thal_7' : 'thal_reversible'}, axis = 'columns', inplace = True)
df['target_true'] = df.target
df.drop(columns = 'target', inplace = True)
df.rename({'target_true': 'target'}, axis = 'columns', inplace = True)
```
```python
df.head()
```
<div>
<style scoped>
.dataframe tbody tr th:only-of-type {
vertical-align: middle;
}
.dataframe tbody tr th {
vertical-align: top;
}
.dataframe thead th {
text-align: right;
}
</style>
<table border="1" class="dataframe">
<thead>
<tr style="text-align: right;">
<th></th>
<th>age</th>
<th>trestbps</th>
<th>chol</th>
<th>thalach</th>
<th>oldpeak</th>
<th>sex_male</th>
<th>cp_atypical</th>
<th>cp_non_anginal</th>
<th>cp_asymptomatic</th>
<th>fbs_true</th>
<th>...</th>
<th>restecg_hypertrophy</th>
<th>exang_yes</th>
<th>slope_flat</th>
<th>slope_downsloping</th>
<th>ca_1</th>
<th>ca_2</th>
<th>ca_3</th>
<th>thal_fixed</th>
<th>thal_reversible</th>
<th>target</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>63</td>
<td>145</td>
<td>233</td>
<td>150</td>
<td>2.3</td>
<td>1</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>1</td>
<td>...</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>1</td>
<td>0</td>
<td>1</td>
</tr>
<tr>
<th>1</th>
<td>37</td>
<td>130</td>
<td>250</td>
<td>187</td>
<td>3.5</td>
<td>1</td>
<td>0</td>
<td>1</td>
<td>0</td>
<td>0</td>
<td>...</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>1</td>
</tr>
<tr>
<th>2</th>
<td>41</td>
<td>130</td>
<td>204</td>
<td>172</td>
<td>1.4</td>
<td>0</td>
<td>1</td>
<td>0</td>
<td>0</td>
<td>0</td>
<td>...</td>
<td>0</td>
<td>0
基于机器学习的冠心病诊断内含数据集和运行环境说明.zip
版权申诉
106 浏览量
2024-05-01
22:05:15
上传
评论
收藏 25.94MB ZIP 举报
基于机器学习的冠心病诊断内含数据集和运行环境说明.zip （180个子文件） 
heart.csv 11KB heart.csv 11KB heart.csv 11KB iris.csv 5KB messy_data.data 1KB tree_gini_Wine_data.gvz 6KB tree_deviance.gvz 1KB tree_entropy.gvz 1KB tree_gini.gvz 1KB tree_gini.gvz 1KB Project 3 - Unsupervised Learning.ipynb 1.74MB Project 1 - Data feature extraction and visualisation.ipynb 618KB Project 2 - Supervised Learning.ipynb 458KB wildfaces_grayscale.mat 11.5MB digits.mat 3.68MB zipdata.mat 3.68MB wine.mat 123KB wine2.mat 117KB synth4.mat 49KB synth5.mat 8KB synth7.mat 8KB synth3.mat 7KB synth1.mat 7KB synth2.mat 7KB xor.mat 7KB body.mat 2KB iris.mat 2KB faithful.mat 2KB synth6.mat 594B README.md 84KB README.md 56KB README.md 32KB README.md 2KB
tree_gini_wine_data.png 3.03MB output_32_0.png 805KB tree_entropy.png 472KB tree_gini.png 451KB output_39_0.png 170KB output_73_0.png 110KB output_42_0.png 99KB output_27_0.png 85KB output_13_0.png 67KB output_37_1.png 60KB output_40_1.png 60KB output_38_1.png 60KB output_39_1.png 59KB output_21_1.png 54KB output_39_2.png 24KB output_40_2.png 24KB output_38_2.png 23KB output_36_0.png 22KB output_37_2.png 21KB output_17_4.png 21KB output_17_3.png 20KB output_17_1.png 20KB output_17_5.png 19KB output_17_2.png 19KB output_69_0.png 18KB output_24_1.png 18KB output_33_1.png 15KB output_30_1.png 14KB output_17_6.png 13KB output_70_1.png 10KB output_64_1.png 10KB output_67_1.png 9KB __init__.py 50KB ANN.py 9KB LinReg Lambda.py 8KB #9 - ANN - binary classification using ANN and logistic regression.py 7KB #2 - ML Data Loading - Cleaning messy data.py 7KB #9 - Regularization - Finding best regularization factor and weights for a linear regression, and compare test errors.py 6KB #9 - ANN - regression using ANN and showing diagram of best neural net.py 6KB #7 - Cross-Validation - Multi-level cross-validation with linear regression and feature selection of attributes.py 5KB #9 - ANN - binary classification using ANN and showing diagram of best neural net.py 4KB #2 - ML Data Loading - Plotting of classification and regression problems.py 4KB #12 - KDE, KNN and ARD - Plotting outliers of hand-written digits using KDE, KNN density and ARD density methods.py 4KB #1 - Basic Python operations - Matrix operations.py 4KB #5 - Data Visualization - Identifying and removing outliers based on data visualization.py 3KB #2 - ML Data Loading - CSV format.py 3KB #3 - PCA - Standardized vs Non-standardized - Cumulative Variance, Projections and Attribute Coefficient - Plots.py 3KB #10 - Boosting - Using AdaBoost technique upon an ensemble of logistic regressions.py 3KB #9 - Regularization - Finding best regularization strength for logistic regression.py 3KB similarity.py 3KB #9 - ANN - multinominal classification using ANN and softmax function.py 3KB #8 - KNN - Classification, plotting and confusion matrix plot for KNN model.py 3KB #7 - Cross-Validation - Model selection based on their mean error difference, using cross-validation.py 3KB #3 - Digit PCA - Visualize digit reconstruction images according to different PCx.py 2KB #7 - Cross-Validation - Test & training error across K cross-validation folds of Decision Tree model.py 2KB #11 - K-means clustering - Cluster images and show their corresponding centroids.py 2KB #12 - GMM - Fitting a GMM and showing its cluster plot.py 2KB #4 - Measures of similarity - Similarity scores in digit recognizing (using all measures!).py 2KB #12 - GMM - Selecting the number of GMM components K using CV, BIC and AIC.py 2KB #5 - Data Visualization - Pairplot showing best dataset class-separation features.py 2KB #8 - Naive Bayes - Multinominal NB error rate, using one-level CV and one-hot encoded categorical variables.py 2KB #9 - Multinominal log regression - Histogram of label classifications and decision boundary plot.py 2KB #4 - Bag-of-words - Creation of document-term matrix.py 2KB #10 - Bagging - Using bagging technique upon an ensemble of logistic regressions.py 2KB #5 - Multivariate Normal Distribution - Generating new handwritten digits using uni and multivariate normal distribution.py 2KB #4 - Bag-of-words - Stemming of document-term matrix.py 2KB #7 - Model error - Misclassification rate, test & training error of Decision Tree model, dependent on model complexity.py 2KB共 180 条
- 1
- 2
资源评论
