Prototype of Decision Support System using Pattern Recognition as an Application of Artificial Intelligence and Machine Learning for Early Diagnosis of Genetic Diseases
With enormous amount of information existing for several diseases, managing and accessing the required data becomes tedious. Moreover, in the healthcare systems, the advent of technology has now proved to be extremely beneficial for storing and retrieving such data. An even more challenging task is the appropriate use of this data that can aid the clinicians, researchers and diagnosticians make an early decision that can help in formulating an appropriate treatment strategy for the patients. Thus, with this background, the present study aimed to design and develop a decision support system using pattern recognition and implementing artificial intelligence (AI) and machine learning (ML) algorithms. As a test case for constructing the model, polymorphisms from twenty cancer exome datasets were utilized for building the model. A basic pattern was recognized for 4181 cancer polymorphisms that were previously identified among various cancer exome datasets. All common nucleotide base substitutions that occurred most frequently were identified. After ruling out other machine learning algorithms in an initial scrutiny to identify the best working algorithm for DSS, a model was designed using supervised random forest technique by selecting relevant features. When polymorphisms from all twenty exome datasets were used for training, prediction for various cancer types were found to be correct. The accuracy of the designed model was found to be 79%. Thus, although the method proposed in the present study provides insights into early detection of various cancer types, the protocol can be applied for early diagnosis of other genetic diseases as well, thereby providing immense support for easy decision-making in healthcare and medicine.
Keywords - Decision Support System, Polymorphisms, Machine Learning, Random Forest, Disease Diagnosis