Mental health is a topic of growing interest in recent years, especially after the COVID-19 pandemic. Its importance and impact on society is very high and fortunately its awareness is also increasing in the last years. Experts have discussed in recent publications the influence that social networks have on people potentially suffering from these illnesses. These studies have concluded that the use of social networks can be a risk factor for the development of an Eating Disorder.
These problems have many symptoms by which they can be diagnosed, such as depression, a nervous mood or by how that person expresses themselves. These are difficult to detect, as they entail a deep and long-lasting professional analysis process, so the diagnosis of an Eating Disorder is not trivial. One possible solution for detection is to use social networks, complementing and modernizing the traditional way in which professionals deal with a patient. These social networks have been used successfully in the detection of mood and mental problems, thus this Master?s Thesis proposes the exploitation of social media data for their use in the detection of Eating Disorders.
We propose the development of a system for the analysis and detection of eating disorder-related mental health problems in social networks. The application will integrate artificial intelligence technologies such as NLP and machine learning to detect these problems working with user data extracted from social media. The purpose of this tool is the detection and analysis of user publications referring to Eating Behavior Disorders on social networks. For this, several subtasks are required:
- Data collection, processing and analysis
- Design and implementation of machine learning models able to detect eating disorders from social media
- Training and evaluation the proposed models using the obtained datasets
- Development of a web application that integrates the proposed models for detecting and analyzing eating disorders
To make all this possible, we will start by analyzing the state of the art to know the existing works on the subject. Afterwards, we will define the system requirements and we will design and develop the system taking into account the information and requirements captured previously. Python programming tools will be used for the design and implementation of the machine learning models where libraries such as NTLK or scikit-learn will be used. Besides, web technologies are going to be used for the application development.