The use of behavioural biometrics, such as movement patterns and keystroke dynamics, in human identity recognition research to strengthen smartphone security is growing. Users usually secure their phones with a PIN or pattern. This paper uses a smartphone keystroke dynamic open dataset with user age information. This Open Dataset is known as the RHU-Keystroke Dynamics dataset. A dataset classification study was conducted utilising the Weighted K-nearest neighbour (W-KNN) method in order to identify the three age categories with the highest accuracy. The four keystroke features that have been collected in this open dataset are used for this classification. The highest average accuracy obtained from this W-KNN method is 83%. The results of the study are explained with a Confusion Matrix diagram and a Receiver Operating Characteristic (ROC) graph. A classification study using this method has successfully increased accuracy and can be utilized in the use of software, as demonstrated in the results of the study. It is expected that future studies will apply other classification methods to keystroke dynamics.
