In the case of preliminary data processing, particularly, in identifying symbols, the determination and categorization of informative symbols or sets of symbols that classify objects is an important issue. Although, there have been a number of methods and algorithms proposed to solve these problems, many more ones are to be solved in this direction. This is due to the fact that most of the proposed approaches are strongly dependent on the nature of the subject of the study, the number of symbols, the type of values that can take symbols, the size of the training sample, and so forth. In addition, there are certain requirements for the above. Besides these, each method or algorithm depends to a large extent on the correctness of the information selection criterion and choosing the appropriate decisive rule that determines the quality of the selected choice. In this regard, the efficiency and reliability of many methods and algorithms are not stable.
The flexible search-based algorithm proposed below is in some sense universal, since symbols can take different types of values, and the proposed informative criterion for selecting a set of symbols is based on minimizing classification errors. In addition, the probability vector used in character selection prevents objects from irrelevant displacement of their important characters out of the selection.
Using this algorithm, the diagnostic data of patients with stenocardia, acute myocardial infarction and cardiac arrhythmia, which are part of ischemic heart disease, were first processed. In this case, the problem of selecting and classifying disease complexes was studied according to the clinical signs and symptoms of patients. Results of the study are presented.
preliminary data processing, criteria for informativeness, character selection, classification error coefficient, flexible random search.
A Transfer Learning-Based Approach with Deep CNN for COVID-19- and Pneumonia-Affected Chest X-ray Image Classification.
Chakraborty S, Paul S, Hasan KMA
SN computer science. 2022; 3(1): 17
Event-Driven Acquisition and Machine-Learning-Based Efficient Prediction of the Li-Ion Battery Capacity.
Mian Qaisar S, AbdelGawad AEE, Srinivasan K
SN computer science. 2022; 3(1): 15
Interrater agreement of contouring of the neurovascular bundles and internal pudendal arteries in neurovascular-sparing magnetic resonance-guided radiotherapy for localized prostate cancer.
Teunissen FR, Wortel RC, Wessels FJ, Claes A, van de Pol SMG, Rasing MJA, Meijer RP, van Melick HHE, de Boer JCJ, Verkooijen HM, van der Voort van Zyp JRN
Clinical and translational radiation oncology. 2022; 32(): 29-34
Diversity Forests: Using Split Sampling to Enable Innovative Complex Split Procedures in Random Forests.
SN computer science. 2022; 3(1): 1
GANBOT: a GAN-based framework for social bot detection.
Najari S, Salehi M, Farahbakhsh R
Social network analysis and mining. 2022; 12(1): 4