Code

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay

### Exploring the dataset.

Code

# Import data and show first 5 rows
df = pd.read_csv("/Users/ShirleyLi/Documents/ANLY501/Lab/m4/Lab4.2/lab4-2-data.csv")
df.head(5)

	Unnamed: 0	X1	age	gender	height	weight	steps	hear_rate	calories	distance	entropy_heart	entropy_setps	resting_heart	corr_heart_steps	norm_heart	intensity_karvonen	sd_norm_heart	steps_times_distance	device	activity
0	1	1	20	1	168.0	65.4	10.771429	78.531302	0.344533	0.008327	6.221612	6.116349	59.0	1.000000	19.531302	0.138520	1.000000	0.089692	apple watch	Lying
1	2	2	20	1	168.0	65.4	11.475325	78.453390	3.287625	0.008896	6.221612	6.116349	59.0	1.000000	19.453390	0.137967	1.000000	0.102088	apple watch	Lying
2	3	3	20	1	168.0	65.4	12.179221	78.540825	9.484000	0.009466	6.221612	6.116349	59.0	1.000000	19.540825	0.138587	1.000000	0.115287	apple watch	Lying
3	4	4	20	1	168.0	65.4	12.883117	78.628260	10.154556	0.010035	6.221612	6.116349	59.0	1.000000	19.628260	0.139208	1.000000	0.129286	apple watch	Lying
4	5	5	20	1	168.0	65.4	13.587013	78.715695	10.825111	0.010605	6.221612	6.116349	59.0	0.982816	19.715695	0.139828	0.241567	0.144088	apple watch	Lying

Code

# Reomonve Unnamed: 0 and X1 columns, convert device to dummy variable, and drop device column.
df = df.drop(columns=['Unnamed: 0', 'X1'])
device_cat = pd.get_dummies(df, columns=['device'])
df = df.drop(columns=['device'])
df

	age	gender	height	weight	steps	hear_rate	calories	distance	entropy_heart	entropy_setps	resting_heart	corr_heart_steps	norm_heart	intensity_karvonen	sd_norm_heart	steps_times_distance	activity
0	20	1	168.0	65.4	10.771429	78.531302	0.344533	0.008327	6.221612	6.116349	59.0	1.000000	19.531302	0.138520	1.000000	0.089692	Lying
1	20	1	168.0	65.4	11.475325	78.453390	3.287625	0.008896	6.221612	6.116349	59.0	1.000000	19.453390	0.137967	1.000000	0.102088	Lying
2	20	1	168.0	65.4	12.179221	78.540825	9.484000	0.009466	6.221612	6.116349	59.0	1.000000	19.540825	0.138587	1.000000	0.115287	Lying
3	20	1	168.0	65.4	12.883117	78.628260	10.154556	0.010035	6.221612	6.116349	59.0	1.000000	19.628260	0.139208	1.000000	0.129286	Lying
4	20	1	168.0	65.4	13.587013	78.715695	10.825111	0.010605	6.221612	6.116349	59.0	0.982816	19.715695	0.139828	0.241567	0.144088	Lying
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
6259	46	0	157.5	71.4	1.000000	35.000000	20.500000	1.000000	0.000000	0.000000	35.0	1.000000	0.000000	0.000000	0.000000	1.000000	Running 7 METs
6260	46	0	157.5	71.4	1.000000	35.000000	20.500000	1.000000	0.000000	0.000000	35.0	1.000000	0.000000	0.000000	1.000000	1.000000	Running 7 METs
6261	46	0	157.5	71.4	1.000000	35.000000	20.500000	1.000000	0.000000	0.000000	35.0	1.000000	0.000000	0.000000	1.000000	1.000000	Running 7 METs
6262	46	0	157.5	71.4	1.000000	35.000000	20.500000	1.000000	0.000000	0.000000	35.0	1.000000	0.000000	0.000000	1.000000	1.000000	Running 7 METs
6263	46	0	157.5	71.4	1.000000	35.000000	20.500000	1.000000	0.000000	0.000000	35.0	1.000000	0.000000	0.000000	1.000000	1.000000	Running 7 METs

6264 rows × 17 columns

### Split the dataset into training and testing sets

Code

# Split the dataframe into X and y and then split X and y into train and test sets.
from sklearn.model_selection import train_test_split

y = df.activity
x = df.drop(columns=['activity'])

X_train, X_test, Y_train, Y_test = train_test_split(x, y, test_size=0.2, random_state=0)

Code

# Show the shape of the train and test sets, and levels of the depencent variable (Y) 
print(X_train.shape)
print(X_test.shape)
print(Y_train.shape)
print(Y_test.shape)

level_y = y.value_counts()
print(level_y)

(5011, 16)
(1253, 16)
(5011,)
(1253,)
Lying             1379
Running 7 METs    1114
Running 5 METs    1002
Running 3 METs     950
Sitting            930
Self Pace walk     889
Name: activity, dtype: int64

### SVM with Linear kernels

Code

# Import svc from sklearn.svm and classsification_report, confusion_matrix from sklearn.metrics.
# Fit the classfier on the training data and predict on the test data. Set the classifier to be linear and C between 0.35-0.75. 
from sklearn.svm import SVC
from sklearn import svm
from sklearn.metrics import classification_report, confusion_matrix

model = svm.LinearSVC(C = 0.55)
model = model.fit(X_train, Y_train)
YP_train = model.predict(X_train)
YP_test = model.predict(X_test)

plt.rcParams['figure.figsize']=10,10
def confusion_plot(y_data,y_pred):
    cm = confusion_matrix(y_data, y_pred)
    disp = ConfusionMatrixDisplay(confusion_matrix=cm)
    disp.plot()
    return plt.show()

confusion_plot(Y_train,YP_train)

/Users/ShirleyLi/opt/anaconda3/lib/python3.9/site-packages/sklearn/svm/_base.py:1206: ConvergenceWarning:

Liblinear failed to converge, increase the number of iterations.

Code

# Calculate the confusion matrix and classification report for the train and test data. 
from sklearn.metrics import classification_report
target_names = ['Lying', 'Running 7 METs', 'Running 5 METs', 'Running 3 METs', 'Sitting', 'Self Pace walk']
train_rep = classification_report(Y_train, YP_train, target_names=target_names)
test_rep = classification_report(Y_test, YP_test, target_names=target_names)
print(train_rep)
print(test_rep)

                precision    recall  f1-score   support

         Lying       0.34      0.36      0.35      1096
Running 7 METs       0.00      0.00      0.00       783
Running 5 METs       0.20      0.58      0.30       809
Running 3 METs       0.50      0.34      0.40       868
       Sitting       0.18      0.06      0.09       703
Self Pace walk       0.22      0.20      0.21       752

      accuracy                           0.27      5011
     macro avg       0.24      0.26      0.23      5011
  weighted avg       0.25      0.27      0.24      5011

                precision    recall  f1-score   support

         Lying       0.31      0.32      0.32       283
Running 7 METs       0.00      0.00      0.00       167
Running 5 METs       0.19      0.59      0.29       193
Running 3 METs       0.57      0.35      0.43       246
       Sitting       0.20      0.05      0.08       186
Self Pace walk       0.19      0.18      0.18       178

      accuracy                           0.26      1253
     macro avg       0.24      0.25      0.22      1253
  weighted avg       0.27      0.26      0.24      1253

Code

# Save the results in a data frame. 
with open('train_rep.csv', 'w') as out:
    out.write(train_rep)
with open('test_rep.csv', 'w') as out:
    out.write(test_rep)
    
train_rep = pd.read_csv('train_rep.csv')
test_rep = pd.read_csv('test_rep.csv')

train_rep = pd.DataFrame(train_rep)
test_rep = pd.DataFrame(test_rep)

Code

# display the results data frame
print(train_rep)
print(test_rep)

                  precision    recall  f1-score   support
0           Lying       0.34      0.36      0.35 ...     
1  Running 7 METs       0.00      0.00      0.00 ...     
2  Running 5 METs       0.20      0.58      0.30 ...     
3  Running 3 METs       0.50      0.34      0.40 ...     
4         Sitting       0.18      0.06      0.09 ...     
5  Self Pace walk       0.22      0.20      0.21 ...     
6        accuracy                           0.27 ...     
7       macro avg       0.24      0.26      0.23 ...     
8    weighted avg       0.25      0.27      0.24 ...     
                  precision    recall  f1-score   support
0           Lying       0.31      0.32      0.32 ...     
1  Running 7 METs       0.00      0.00      0.00 ...     
2  Running 5 METs       0.19      0.59      0.29 ...     
3  Running 3 METs       0.57      0.35      0.43 ...     
4         Sitting       0.20      0.05      0.08 ...     
5  Self Pace walk       0.19      0.18      0.18 ...     
6        accuracy                           0.26 ...     
7       macro avg       0.24      0.25      0.22 ...     
8    weighted avg       0.27      0.26      0.24 ...     

Code

# Display Confusion Matrix for the test data. Remember to use the ConfusionMatrixDisplay function.
confusion_plot(Y_test,YP_test)

## SVM with Polynomial kernels

Code

# Import svc from sklearn.svm and classsification_report, confusion_matrix from sklearn.metrics.
# Fit the classfier on the training data and predict on the test data. Set the classifier to be polynomial, C between 0.35-0.75, and degree = 2. 
model = svm.SVC(C = 0.55, kernel = 'poly',degree = 2)
model = model.fit(X_train, Y_train)
YP_train = model.predict(X_train)
YP_test = model.predict(X_test)

Code

# Calculate the confusion matrix and classification report for the train and test data. 
train_rep = classification_report(Y_train, YP_train, target_names=target_names)
test_rep = classification_report(Y_test, YP_test, target_names=target_names)
print(train_rep)
confusion_plot(Y_train,YP_train)
print(test_rep)
confusion_plot(Y_test,YP_test)

/Users/ShirleyLi/opt/anaconda3/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning:

Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.

/Users/ShirleyLi/opt/anaconda3/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning:

Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.

/Users/ShirleyLi/opt/anaconda3/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning:

Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.

/Users/ShirleyLi/opt/anaconda3/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning:

Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.

/Users/ShirleyLi/opt/anaconda3/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning:

Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.

/Users/ShirleyLi/opt/anaconda3/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1318: UndefinedMetricWarning:

Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.

                precision    recall  f1-score   support

         Lying       0.22      1.00      0.36      1096
Running 7 METs       0.31      0.01      0.02       783
Running 5 METs       0.00      0.00      0.00       809
Running 3 METs       0.00      0.00      0.00       868
       Sitting       0.00      0.00      0.00       703
Self Pace walk       0.00      0.00      0.00       752

      accuracy                           0.22      5011
     macro avg       0.09      0.17      0.06      5011
  weighted avg       0.10      0.22      0.08      5011

                precision    recall  f1-score   support

         Lying       0.22      0.99      0.37       283
Running 7 METs       0.00      0.00      0.00       167
Running 5 METs       0.00      0.00      0.00       193
Running 3 METs       0.00      0.00      0.00       246
       Sitting       0.00      0.00      0.00       186
Self Pace walk       0.00      0.00      0.00       178

      accuracy                           0.22      1253
     macro avg       0.04      0.16      0.06      1253
  weighted avg       0.05      0.22      0.08      1253

Code

# Save the results in a data frame.
with open('train_rep2.csv', 'w') as out:
    out.write(train_rep)
with open('test_rep2.csv', 'w') as out:
    out.write(test_rep)
    
train_rep = pd.read_csv('train_rep2.csv')
test_rep = pd.read_csv('test_rep2.csv')

train_rep = pd.DataFrame(train_rep)
test_rep = pd.DataFrame(test_rep)

Code

# display the results data frame
print(train_rep)
print(test_rep)

                  precision    recall  f1-score   support
0           Lying       0.22      1.00      0.36 ...     
1  Running 7 METs       0.31      0.01      0.02 ...     
2  Running 5 METs       0.00      0.00      0.00 ...     
3  Running 3 METs       0.00      0.00      0.00 ...     
4         Sitting       0.00      0.00      0.00 ...     
5  Self Pace walk       0.00      0.00      0.00 ...     
6        accuracy                           0.22 ...     
7       macro avg       0.09      0.17      0.06 ...     
8    weighted avg       0.10      0.22      0.08 ...     
                  precision    recall  f1-score   support
0           Lying       0.22      0.99      0.37 ...     
1  Running 7 METs       0.00      0.00      0.00 ...     
2  Running 5 METs       0.00      0.00      0.00 ...     
3  Running 3 METs       0.00      0.00      0.00 ...     
4         Sitting       0.00      0.00      0.00 ...     
5  Self Pace walk       0.00      0.00      0.00 ...     
6        accuracy                           0.22 ...     
7       macro avg       0.04      0.16      0.06 ...     
8    weighted avg       0.05      0.22      0.08 ...     

Code

# Display Confusion Matrix for the test data. Remember to use the ConfusionMatrixDisplay function.
confusion_plot(Y_test,YP_test)

## SVM with RBF kernels

Code

# Import svc from sklearn.svm and classsification_report, confusion_matrix from sklearn.metrics.
# Fit the classfier on the training data and predict on the test data. Set the classifier to be linear and C between 0.35-0.75. 
model = svm.SVC(C = 0.55, kernel = 'rbf')
model = model.fit(X_train, Y_train)
YP_train = model.predict(X_train)
YP_test = model.predict(X_test)

Code

# Calculate the confusion matrix and classification report for the train and test data. 
train_rep = classification_report(Y_train, YP_train, target_names=target_names)
test_rep = classification_report(Y_test, YP_test, target_names=target_names)
print(train_rep)
confusion_plot(Y_train,YP_train)
print(test_rep)
confusion_plot(Y_test,YP_test)

                precision    recall  f1-score   support

         Lying       0.22      0.97      0.36      1096
Running 7 METs       0.55      0.01      0.02       783
Running 5 METs       0.67      0.00      0.00       809
Running 3 METs       0.30      0.08      0.12       868
       Sitting       0.47      0.01      0.02       703
Self Pace walk       0.17      0.00      0.01       752

      accuracy                           0.23      5011
     macro avg       0.40      0.18      0.09      5011
  weighted avg       0.39      0.23      0.11      5011

                precision    recall  f1-score   support

         Lying       0.23      0.94      0.36       283
Running 7 METs       0.00      0.00      0.00       167
Running 5 METs       0.00      0.00      0.00       193
Running 3 METs       0.30      0.07      0.11       246
       Sitting       0.60      0.02      0.03       186
Self Pace walk       0.00      0.00      0.00       178

      accuracy                           0.23      1253
     macro avg       0.19      0.17      0.08      1253
  weighted avg       0.20      0.23      0.11      1253

Code

# Save the results in a data frame.
with open('train_rep3.csv', 'w') as out:
    out.write(train_rep)
with open('test_rep3.csv', 'w') as out:
    out.write(test_rep)
    
train_rep = pd.read_csv('train_rep3.csv')
test_rep = pd.read_csv('test_rep3.csv')

train_rep = pd.DataFrame(train_rep)
test_rep = pd.DataFrame(test_rep)

Code

# display the results data frame
print(train_rep)
print(test_rep)

                  precision    recall  f1-score   support
0           Lying       0.22      0.97      0.36 ...     
1  Running 7 METs       0.55      0.01      0.02 ...     
2  Running 5 METs       0.67      0.00      0.00 ...     
3  Running 3 METs       0.30      0.08      0.12 ...     
4         Sitting       0.47      0.01      0.02 ...     
5  Self Pace walk       0.17      0.00      0.01 ...     
6        accuracy                           0.23 ...     
7       macro avg       0.40      0.18      0.09 ...     
8    weighted avg       0.39      0.23      0.11 ...     
                  precision    recall  f1-score   support
0           Lying       0.23      0.94      0.36 ...     
1  Running 7 METs       0.00      0.00      0.00 ...     
2  Running 5 METs       0.00      0.00      0.00 ...     
3  Running 3 METs       0.30      0.07      0.11 ...     
4         Sitting       0.60      0.02      0.03 ...     
5  Self Pace walk       0.00      0.00      0.00 ...     
6        accuracy                           0.23 ...     
7       macro avg       0.19      0.17      0.08 ...     
8    weighted avg       0.20      0.23      0.11 ...     

Code

# Display Confusion Matrix for the test data. Remember to use the ConfusionMatrixDisplay function.
confusion_plot(Y_test,YP_test)

## SVM with Sigmoid kernels

Code

# Import svc from sklearn.svm and classsification_report, confusion_matrix from sklearn.metrics.
# Fit the classfier on the training data and predict on the test data. Set the classifier to be linear and C between 0.35-0.75. 
model = svm.SVC(C = 0.55, kernel = 'sigmoid',degree = 2)
model = model.fit(X_train, Y_train)
YP_train = model.predict(X_train)
YP_test = model.predict(X_test)

Code

# Calculate the confusion matrix and classification report for the train and test data. 
train_rep = classification_report(Y_train, YP_train, target_names=target_names)
test_rep = classification_report(Y_test, YP_test, target_names=target_names)
print(train_rep)
confusion_plot(Y_train,YP_train)
print(test_rep)
confusion_plot(Y_test,YP_test)

                precision    recall  f1-score   support

         Lying       0.21      0.54      0.30      1096
Running 7 METs       0.15      0.01      0.01       783
Running 5 METs       0.18      0.38      0.25       809
Running 3 METs       0.19      0.08      0.11       868
       Sitting       0.18      0.02      0.04       703
Self Pace walk       0.05      0.00      0.00       752

      accuracy                           0.20      5011
     macro avg       0.16      0.17      0.12      5011
  weighted avg       0.17      0.20      0.13      5011

                precision    recall  f1-score   support

         Lying       0.19      0.47      0.27       283
Running 7 METs       0.14      0.01      0.01       167
Running 5 METs       0.15      0.35      0.21       193
Running 3 METs       0.17      0.06      0.09       246
       Sitting       0.18      0.03      0.05       186
Self Pace walk       0.14      0.01      0.01       178

      accuracy                           0.18      1253
     macro avg       0.16      0.15      0.11      1253
  weighted avg       0.17      0.18      0.12      1253

Code

# Save the results in a data frame.
with open('train_rep4.csv', 'w') as out:
    out.write(train_rep)
with open('test_rep4.csv', 'w') as out:
    out.write(test_rep)
    
train_rep = pd.read_csv('train_rep4.csv')
test_rep = pd.read_csv('test_rep4.csv')

train_rep = pd.DataFrame(train_rep)
test_rep = pd.DataFrame(test_rep)

Code

# display the results data frame
print(train_rep)
print(test_rep)

                  precision    recall  f1-score   support
0           Lying       0.21      0.54      0.30 ...     
1  Running 7 METs       0.15      0.01      0.01 ...     
2  Running 5 METs       0.18      0.38      0.25 ...     
3  Running 3 METs       0.19      0.08      0.11 ...     
4         Sitting       0.18      0.02      0.04 ...     
5  Self Pace walk       0.05      0.00      0.00 ...     
6        accuracy                           0.20 ...     
7       macro avg       0.16      0.17      0.12 ...     
8    weighted avg       0.17      0.20      0.13 ...     
                  precision    recall  f1-score   support
0           Lying       0.19      0.47      0.27 ...     
1  Running 7 METs       0.14      0.01      0.01 ...     
2  Running 5 METs       0.15      0.35      0.21 ...     
3  Running 3 METs       0.17      0.06      0.09 ...     
4         Sitting       0.18      0.03      0.05 ...     
5  Self Pace walk       0.14      0.01      0.01 ...     
6        accuracy                           0.18 ...     
7       macro avg       0.16      0.15      0.11 ...     
8    weighted avg       0.17      0.18      0.12 ...     

Code

# Display Confusion Matrix for the test data. Remember to use the ConfusionMatrixDisplay function.
confusion_plot(Y_test,YP_test)