#MachineLearning #SupervisedLearning #Classification
By Billy Gustave
Run\Walk Classifier ¶
Goal:
- Classify activities as walk or run
- Data: run_or_walk.csv
- {0:'Walk', 1:'Run'}
- Naive Bayes Classifier
- Analyzing accuracy of acceleration vs gyro vs all
Data Exploration ¶
In [1]:
import numpy as np, pandas as pd, matplotlib.pyplot as plt, seaborn as sns
In [2]:
df = pd.read_csv('run_or_walk.csv')
df.shape
Out[2]:
In [3]:
df.head()
Out[3]:
In [4]:
df.describe()
Out[4]:
Data Cleaning ¶
In [5]:
df.info()
No Missing Values
In [6]:
# Features and Target
X = df.drop(['date','time','username','wrist','activity'], axis=1)
y = df.activity
X.shape
Out[6]:
In [7]:
from sklearn.model_selection import train_test_split
# train/test split
x_train,x_test,y_train,y_test = train_test_split(X,y,test_size=0.2,random_state=1)
In [8]:
fig, ax = plt.subplots(figsize=(16,14))
sns.heatmap(x_train.corr(), cmap='Reds', annot=True, linewidths=.5, ax=ax)
Out[8]:
No highly correlated features
Modeling ¶
Using Kfold and Cross Validation:
In [9]:
from sklearn.model_selection import cross_val_score, KFold
kfold = KFold(n_splits=10, random_state=1, shuffle=True)
In [10]:
from sklearn.naive_bayes import GaussianNB
nbc = GaussianNB()
acc_features = ['acceleration_x', 'acceleration_y', 'acceleration_z']
gyro_features = ['gyro_x', 'gyro_y', 'gyro_z']
# Calculate accuracy scores
all_score = cross_val_score(nbc, x_train, y_train, cv=kfold, scoring='accuracy').mean()
print("Accuracy : {} ".format(all_score))
acc_score = cross_val_score(nbc, x_train[acc_features], y_train, cv=kfold, scoring='accuracy').mean()
print("Acceleration accuracy : {} ".format(acc_score))
gyro_score = cross_val_score(nbc, x_train[gyro_features], y_train, cv=kfold, scoring='accuracy').mean()
print("Gyro accuracy : {} ".format(gyro_score))
Testing
In [11]:
nbc = GaussianNB()
all_pred_y = nbc.fit(x_train,y_train).predict(x_test)
acc_pred_y = nbc.fit(x_train[acc_features],y_train).predict(x_test[acc_features])
gyro_pred_y = nbc.fit(x_train[gyro_features],y_train).predict(x_test[gyro_features])
In [12]:
from sklearn.metrics import accuracy_score, confusion_matrix, classification_report
All
In [13]:
accuracy_score(all_pred_y, y_test)
Out[13]:
In [14]:
confusion_matrix(y_test, all_pred_y)
Out[14]:
In [15]:
print(classification_report(y_test,all_pred_y))
Acceleration
In [16]:
accuracy_score(acc_pred_y, y_test)
Out[16]:
In [17]:
confusion_matrix(y_test, acc_pred_y)
Out[17]:
In [18]:
print(classification_report(y_test,acc_pred_y))
Gyro
In [19]:
accuracy_score(gyro_pred_y, y_test)
Out[19]:
In [20]:
confusion_matrix(y_test, gyro_pred_y)
Out[20]:
In [21]:
print(classification_report(y_test,gyro_pred_y))
Acceleration is the determining factor for classifying running or walking