Human Activity Recognition using Smartphone Dataset- ML Python

Hey ML Enthusiasts, In this article, we are going to create a Human Activity Recognition Model using Machine Learning in Python. Before Proceeding further in the article, you are advised to download Dataset and human-activity-recognition (Notebook)

The Dataset contains various sensors data, related to various activities performed by different individuals.

Requirements:

Keras
Python == 3.6

Code Overview: Human Activity Recognition using Smartphone Dataset in Python

# Lets load Train CSV
df_train = pd.read_csv('https://q2z5x2y2.rocketcdn.me/kaggle/input/human-activity-recognition-with-smartphones/train.csv')
df_train.head()

We are going to load the data frames and then do Feature Engineering.

We will separate features and Labels:

x_train = df_train.iloc[:,0:-2]
x_train = np.array(x_train)
x_train.shape

And now will encode the Labels into 0 and 1 format:

from sklearn.preprocessing import LabelEncoder
lb = LabelEncoder()
y_train = lb.fit_transform(y_train)

# Lets encode this 
from keras.utils import to_categorical
y_train = to_categorical(y_train)
y_train

Using TensorFlow backend.

array([[0., 0., 1., 0., 0., 0.],
       [0., 0., 1., 0., 0., 0.],
       [0., 0., 1., 0., 0., 0.],
       ...,
       [0., 0., 0., 0., 0., 1.],
       [0., 0., 0., 0., 0., 1.],
       [0., 0., 0., 0., 0., 1.]])

Now we will create our model:

# Lets Prepare up the model
model = Sequential()
model.add(Dense(256,input_shape=(x_train.shape[1],1)))
model.add(Dense(128))
model.add(Dropout(0.2))
model.add(Dense(256))
model.add(Dense(128))
model.add(Flatten())
model.add(Dense(y_train.shape[1]))
model.add(Activation('softmax'))
model.summary()

Model: "sequential_1"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense_1 (Dense)              (None, 561, 256)          512       
_________________________________________________________________
dense_2 (Dense)              (None, 561, 128)          32896     
_________________________________________________________________
dropout_1 (Dropout)          (None, 561, 128)          0         
_________________________________________________________________
dense_3 (Dense)              (None, 561, 256)          33024     
_________________________________________________________________
dense_4 (Dense)              (None, 561, 128)          32896     
_________________________________________________________________
flatten_1 (Flatten)          (None, 71808)             0         
_________________________________________________________________
dense_5 (Dense)              (None, 6)                 430854    
_________________________________________________________________
activation_1 (Activation)    (None, 6)                 0         
=================================================================
Total params: 530,182
Trainable params: 530,182
Non-trainable params: 0

We would train our model and found a 95% accuracy.

Congrats…We have successfully built the Human Activity Recognition model.

If you face any problem, drop down your feedback in the comments section

For More Projects, See Language Translator (RNN BiDirectional LSTMs and Attention) in Python.

Human Activity Recognition using Smartphone Dataset- ML Python

Requirements:

Code Overview: Human Activity Recognition using Smartphone Dataset in Python

Leave a Reply Cancel reply

Related Posts