This paper introduces a new method for automatically predicting the response time of the Internet users to solve the CAPTCHA test by a regression tree strategy. The input to the model is a set of demographic features of the user: (i) age, (ii) education level, and (iii) Internet experience level. The experiment is performed on 114 Internet users who are invited to solve three image and interactive CAPTCHA tests: (i) FunCAPTCHA, (ii) home numbers, and (iii) picture of the CAPTCHA. Collected demographic features and response time for each user are processed by the regression tree approach in order to evaluate the prediction accuracy. Obtained results revealed the ability of the model in correctly predicting the response time of the specific CAPTCHA types. This represents an invaluable analysis in the state-of-the-art for designing new CAPTCHA types which are more accustomed to specific categories of Internet users.
File in questo prodotto:
Non ci sono file associati a questo prodotto.