Quantum deep learning in Parkinson’s disease prediction using hybrid quantum–classical convolution neural network

Abstract

Deep learning, also known as DL, holds great potential within the field of artificial intelligence. Fast problem-solving approaches are widely used in quantum computing. Large multidimensional space is utilized to categorize and address intricate problems. The different algorithms have the ability to interact in a space with multiple dimensions and find solutions to the problems. Quantum deep learning facilitates different mining procedures by incorporating precise advancements in quantum computing. Prompt and accurate identification during the early stages of progression is crucial for various severe and life-threatening illnesses like cancer, hepatotoxicity, cardio toxicity, nephrotoxicity, and others. Currently, there is a critical need to create rapid, precise, and highly effective approaches for predicting different diseases. These methods should also be feasible and nonintrusive. Dementia, a highly hazardous condition, has a significant impact on the human nervous system. Dementia often includes Parkinson’s as one of its prominent symptoms. The patient’s entire operational behavior will be impacted. The proposed system is utilizing machine learning and quantum computing to develop a method for predicting Parkinson’s disease based on speech signals. Quantum computers can be used to assist in identifying cancer by using a hybrid quantum–classical convolution neural network (QCCNN). This network is inspired by convolution neural networks (CNNs) but has been modified for quantum computing in order to improve the process of mapping features. Dimensionality reduction algorithms, principal component analysis (PCA) are applied to the preprocessed dataset to make predictions about diseases. The standard dataset from UCI machine learning repository will be used to determine the performance of the model. Ensemble models exceed the precision of highly accurate techniques such as neural networks. To demonstrate the superior detection capability of our model, we have compared its performance with several advanced machine learning and deep learning-based methods for Parkinson’s disease detection.