/Mental-Heath-Disease-Detection

Primary LanguageJupyter Notebook

Mental-Heath-Disease-Detection

Keywords: Tensorflow, keras, numpy, pandas, CNN, Reddit, matplotlib, seaborn , Jupyter Notebook.

Installation

  1. Clone this Repository
git clone https://github.com/ankita2002/Mental-Heath-Disease-Detection.git
  1. Open Jupyter Notebook and Run the files or Directly use the pickle files.

Approach

The Project is built to Indentify Mental health Diseases from text.
The Diseases currently focused are Depression, Bipolar Disorder and Anxiety.

Steps for Model Building.
  1. Web scrapped 1 lakh+ real world data from reddit posts forum of Depression, Anxiety and Bipolar Disorder.
  2. Lemmatization and Tokenization of Text.
  3. Feature Selection: Identification of Unique tokens.
  4. Spliting the Training and testing data in the ratio of 0.8: 0.2.
  5. Using CNN to train the model. (Using embedding and Maxpooling).
  6. Calculate the loss function using Binary Cross Entropy and Adam Optimizer for Optimization.