Gender determination of given handwriting is crucial in various applications. In forensic, the reliable determination of the gender of handwriting helps forensic document examiners to narrow down the suspects at hand. However, high accuracy rates are needed to eliminate the possible errors. To this end, handwriting samples including 11 sentences containing the letters “b, d, f, g, h, k, t, y, p” at initial, medial, and end positions were collected from 50 female and 50 male participants. The ascender and descender parts of these letters were measured in millimeters, and statistically significant differences were found between the handwriting of the two genders. To increase the accuracy rates, Logistics Regression (LR), k-Nearest Neighbors (k-NN), Support Vector Machine (SVM), and Artificial Neural Network (ANN), the four most successful methods in machine learning studies were applied to the data and accuracy rates of 0.65, 0.60, 0.71, and 0.82 respectively. The obtained accuracy rates are relatively high with the methods used in the current study. With the addition of parameters such as age and hand in handwriting determination to gender parameter in further studies, it is believed that higher accuracy rates can be achieved in the determination of the owner of the handwriting in question with the model applied in this study.

Gender determination of given handwriting is crucial in various applications. In forensic, the reliable determination of the gender of handwriting helps forensic document examiners to narrow down the suspects at hand. However, high accuracy rates are needed to eliminate the possible errors. To this end, handwriting samples including 11 sentences containing the letters “b, d, f, g, h, k, t, y, p” at initial, medial, and end positions were collected from 50 female and 50 male participants. The ascender and descender parts of these letters were measured in millimeters, and statistically significant differences were found between the handwriting of the two genders. To increase the accuracy rates, Logistics Regression (LR), k-Nearest Neighbors (k-NN), Support Vector Machine (SVM), and Artificial Neural Network (ANN), the four most successful methods in machine learning studies were applied to the data and accuracy rates of 0.65, 0.60, 0.71, and 0.82 respectively. The obtained accuracy rates are relatively high with the methods used in the current study. With the addition of parameters such as age and hand in handwriting determination to gender parameter in further studies, it is believed that higher accuracy rates can be achieved in the determination of the owner of the handwriting in question with the model applied in this study.

No abstract