An Application of a Hybrid Intelligent System for Diagnosing Primary Headaches

Abstract

(1) Background: Modern medicine generates a great deal of information that stored in medical databases. Simultaneously, extracting useful knowledge and making scientific decisions for diagnosis and treatment of diseases becomes increasingly necessary. Headache disorders are the most prevalent of all the neurological conditions. Headaches have not only medical but also great socioeconomic significance. The aim of this research is to develop an intelligent system for diagnosing primary headache disorders. (2) Methods: This research applied various mathematical, statistical and artificial intelligence techniques, among which the most important are: Calinski-Harabasz index, Analytical Hierarchy Process, and Weighted Fuzzy C-means Clustering Algorithm. These methods, techniques and methodologies are used to create a hybrid intelligent system for diagnosing primary headache disorders. The proposed intelligent diagnostic system is tested with original real-world data set with different metrics. (3) Results: First at all, nine of 20 attributes – features from International Headache Society (IHS) criteria are selected, and then only five most important attributes from IHS criteria are selected. The calculation result based on the Calinski–Harabasz index value (178) for the optimal number of clusters is three, and they present three classes of headaches: (i) migraine, (ii) tension-type headaches (TTHs), and (iii) other primary headaches (OPHs). The proposed hybrid intelligent system shows the following quality metrics: Accuracy 75%; Precision 67% for migraine, 74% for TTHs, 86% for OPHs, and Average Precision 77%; Recall 86% for migraine, 73% for TTHs, 67% for OPHs, Average Recall 75%; F1 score 75% for migraine, 74% for TTHs, 75% for OPHs, and Average F1 score 75%. (4) Conclusions: The hybrid intelligent system presents qualitative and respectable experimental results. The implementation of existing diagnostics systems and the development of new diagnostics systems in medicine is necessary in order to help physicians make quality diagnosis and decide the best treatments for the patients.

Verification of a clinical decision support system for the diagnosis of headache disorders based on patient–computer interactions: a multi-center study

“…Considering the incompleteness of the language rules used when human experts express knowledge, a study used the Learning-From-Examples algorithm to improve correct recognition rate [ 20 ]. Another study developed a hybrid intelligent system for diagnosing primary headache disorders, applying various mathematical, statistical, and artificial intelligence techniques [ 21 ]. Although various tools have been developed to improve the diagnostic algorithm, no computer-assisted systems have obtained headache information directly from the patient.…”

The Clinical Relevance of Artificial Intelligence in Migraine

“…Furthermore, it is possible to use hybrid intelligent approaches [28] or ML decision trees to realise systems that are able to make a reliable diagnosis of primary or secondary headache based on ICH-3 criteria [1]. The results of a cross-sectional study on 202 patients showed how a Computer-based Diagnostic Engine (CDE) proved to be a reliable tool compared to a semi-structured interview performed by a headache specialist.…”

“…Such classification could improve patient management, influencing the physician's approach based on the main pain pattern reported by the patient. Simic, 2021 [28] Various mathematical, statistical, and artificial intelligence techniques, including decision-making methodology and clustering methods…”

Cluster Predictive Model Using Affinity Propagation Algorithm to Group Mushroom 5.8s rRNA Sequences

“…The author of this research developed a hybrid intelligent system for diagnosing primary headaches. In their path of research they incorporated various methods like Calinski-Harabasz, Analytical hierarchy process, and weighted fuzy-Cmeans algorithms to give qualitative and expected experiment outcomes that help to make quality decisions for further treatments (16) . The research work carried out to identify the optimal number of clusters has used 21 different datasets which have selective features to demonstrate the similarity of each other.…”