Improving the Quality of Life for Intellectually Disabled Elderly People using Artificial Intelligence Techniques

Identificador:  TDX:4336

Handle: https://hdl.handle.net/20.500.11797/TDX4336

Autors:  Yadav, Gaurav Kumar

Resum:
This thesis work is divided into two parts. The first part proposes to predict future human motion by observing past motion. We use the inception residual blocks to learn the temporal feature and the Convolutional Graph Network to learn the spatial features. Using this method, we can produce future poses that are more precise than those produced by other cutting-edge methods. Further, we empower our human motion prediction model to predict using even out-ofdistribution (OoD) data by augmenting discriminative and generative models with regularisation using linear matrices. This proposed network has shown much-improved results for both out-of-distribution and in-distribution scenarios. In the work's second part, we move towards improving the quality of life (QoL) of dependent individuals with intellectual disabilities in guardianship entities with the help of artificial intelligence techniques. More specifically, we proposed that our ML-based models are trained and tested to predict the index value, standard scale value, and support intensity scale value related to the QoL. We evaluate each of the three models using various evaluation metrics of different ML algorithms calculating each task's performance. Subsequently, the proposed method generates a support report containing the required actions to improve the deficit dimensions and quality of life. It helps professionals to track the patient's progress. Further, we use machine learning-based models to predict the values of eight dimensions from the sixty-nine questionnaire responses.