Scribd
Upload
7 views9 pages

ML Model for Depression & Anxiety Prediction

This document describes a machine learning model using convolutional neural networks to predict depression, anxiety, and stress levels during the COVID-19 pandemic. The model uses emotion recognition from facial expressions, audio, text, and physiological data to classify mental health states. A convolutional neural network is employed which applies filters across image layers to detect features, followed by pooling and fully connected layers to compile the extracted data into probabilistic classifications of mental health conditions.

Uploaded by

Rajesh Ece
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
7 views9 pages

ML Model for Depression & Anxiety Prediction

This document describes a machine learning model using convolutional neural networks to predict depression, anxiety, and stress levels during the COVID-19 pandemic. The model uses emotion recognition from facial expressions, audio, text, and physiological data to classify mental health states. A convolutional neural network is employed which applies filters across image layers to detect features, followed by pooling and fully connected layers to compile the extracted data into probabilistic classifications of mental health conditions.

Uploaded by

Rajesh Ece
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd

Machine Learning based Depression,

Anxiety, and Stress Predictive Model during


COVID-19 Crisis
Outline

 Introduction

 Methodology

 Deep learning

 Convolutional Neural Network

 Conclusion
Introduction

 Crisis – extremely stressful event – ambiguous, uncertain, and unexpected


conditions
 Psychological responses – Anxiety, anger, emotional distress, depression,
uncertainty and sense of powerlessness
 High levels of stress and fear – social distancing, lockdown and increasing rates of
new cases and mortalities
Methodology

 Emotion recognition is the process of identifying human emotion


‣ recognition of facial expressions from video
‣ spoken expressions from audio
‣ written expressions from text, and
‣ physiology as measured by wearables
 Security and healthcare purposes
Deep Learning

 Deep learning – a type of supervised machine learning.


 Perform classification tasks directly from images, text, or sound.
 Deep learning is usually implemented using a neural network.
 “Deep” – number of layers in the network – the more the layers, the deeper the
network.
Convolutional Neural Network - CNN

 Hundreds of layers and each layer learn to detect different features of an image.
 Filters are applied to each training image at different resolutions and size, and the
output of each convolved image is used as the input to the next layer.
1. Image Input Layer:
 Input of the whole CNN
 Pixel matrix of the image
2. Convolutional Layer:
 Major building blocks of CNN
 Transforms the input image – to extract features
 The image is convolved with a kernel (or filter)
 Convolved features – stacked one after the other – to create an output.
3. Batch Normalization Layer:
 Make artificial neural networks – faster and more stable
 Normalization of inputs – re-centering and re-scaling
4. ReLU Layer:
 ReLU – Rectified Linear Unit
 Filtering information that propagates forward through the network
 Elementwise operation on the input
 If the input is negative, then it is made zero.
5. Max Pooling Layer:
 A pooling operation – selects the maximum element from the region of the
feature map
 Output – a feature map containing the most prominent features of the previous
feature map
6. Fully connected Layer:
 All the inputs from one layer are connected to every activation unit of the next
layer
 Compiles the data extracted by previous layers to form the final output
7. SoftMax Layer:
 Implemented just before the output layer.
 Converts output of the last layer in into a probability distribution.
8. Classification Layer:
 Computes the cross-entropy loss for classification and weighted classification
tasks with mutually exclusive classes
 Infers the number of classes from the output size of the previous layer

You might also like