Abstract:

Recently, identity recognition mainly includes issues such as age recognition and gender recognition both in academic and other fields (industrial, computer, health, etc.). It is a commonly researched subject. Gender recognition (CT) is determining the individual’s gender on the basis of the characteristics that distinguish between men and women. In the field of artificial intelligence, CT is considered one of the most important applications of the sample recognition method. In this study, three (3) different characteristics extraction methods for gender recognition (CT) were proposed using the signs obtained from the vibration measurements, magnetometers and gyroscopic sensors, which were embedded in 5 different areas of individuals. Extrusion of authenticity from signs is one of the most important stages of CT. Because the success of the CT depends on the subjects. However, getting out the proper qualities for CT is a difficult problem. Signals obtained from the sensors have been applied different conversion methods, including a Dimensional Local Twin Examples (1B-YÖ), a Dimensional Solid Local Twin Examples (1B-YÖ) and a Weighted Dimensional Solid Local Twin Examples (A-1B-YÖ). After the conversion processes, new signs have been obtained statistical properties. Classification processes have been carried out using different machine learning methods (SVM, RF, YSA, Knn) using these properties. According to the results obtained, 1B-YÖ (96.04), 1B-YÖ (96.72) and A-1B-YÖ (97.28) methods have been shown to provide effective properties for CT. Thanks to the new approaches proposed in this study, it has been determined that CT processing is performed with a high success rate using sensor signals.

Keywords: