100
Internet of Things (IoT)
6.1 Introduction
Knowledge-based user authentication technique is a common and easy access control mechanism. But people are uninspired while choosing a healthy password or PIN.
It increases the probability of guessing attacks. In this situation, to minimize these attacks,
keystroke dynamics is a good choice; here users are not only identified by the password
but their typing style is also accounted for. Keystroke dynamics is the method of analyzing
typing pattern on a computer keyboard or touch screen and classifying the users based
on their regular typing rhythm. It is a behavioral biometric characteristic which we have
learned in our life and relates to the issues in human identification/authentication. This is
the method where people can be identified by their typing style similar to hand writing or
voice print. Being noninvasive and cost-effective, this method is a good field of research.
But the performance of keystroke dynamics is less than other popular morphological
biometric characteristics like face print, iris, and finger print recognition due to high rate
of intraclass variation or high Failure to Enroll Rate (FER). So, this technique demands
higher level of security. In this chapter, we are interested in investigating the integration
of the soft biometric features, gender and age group, with the existing keystroke dynamics
user authentication systems proposed by [1–3].
We have investigated the probability of predicting gender and age group based on typing
pattern. This is possible if all the patterns of certain peer are similar and dissimilar from
one peer to another. As per the research direction, it is possible to predict the gender with
88.55% to 95.04% accuracy based on typing pattern on the keyboard. Similarly, we also
obtained 84.75% accuracy with regard to gender based on typing pattern on touch screen.
If we use only gender information as an extra feature, then we can achieve 3.5% to 7.72%
of accuracy. Similarly, the age group (18–30/30+ years) can also be predicted based on
typing pattern. We obtained 86.87% to 94.68% accuracy with regard to prediction of age
group (18–30/30+ years) based on typing pattern on keyboard, whereas 84.75% of accuracy
was obtained by analyzing the typing pattern on touch screen with regard to age group
(7–29/30–65 years). We also analyzed the age group (≤18/18+ years) and obtained 89.2% to
92% accuracy based on typing pattern on touch screen.
These two biometric traits have low discriminating power but can be used as additional
soft biometric features to reduce the error rate in keystroke dynamics user authentication
systems. This technique can also be used in e-commerce sites to reach out to the right
client to avoid adverse products more efficiently based on the gender and age group.
There are many application areas where automatic gender and age group identification
methods can be applied like any surveillance system, online automatic user account
profiling, and protection of minors from online threats.
The main objective of this study is to develop a model that can identify the gender and
age group of users through the way of typing on a computer keyboard and touching a
computer screen for a predefined text, and it increases the accuracy by recognizing this
soft biometric information as additional features in keystroke dynamics user authentication systems.
The major goal and contributions of this chapter are as follows:
• Develop an efficient model to recognize the gender and age group automatically
from typing pattern.
• Validate the model on keystroke dynamics dataset collected through a touch
screen device.
Internet of Things (IoT)
6.1 Introduction
Knowledge-based user authentication technique is a common and easy access control mechanism. But people are uninspired while choosing a healthy password or PIN.
It increases the probability of guessing attacks. In this situation, to minimize these attacks,
keystroke dynamics is a good choice; here users are not only identified by the password
but their typing style is also accounted for. Keystroke dynamics is the method of analyzing
typing pattern on a computer keyboard or touch screen and classifying the users based
on their regular typing rhythm. It is a behavioral biometric characteristic which we have
learned in our life and relates to the issues in human identification/authentication. This is
the method where people can be identified by their typing style similar to hand writing or
voice print. Being noninvasive and cost-effective, this method is a good field of research.
But the performance of keystroke dynamics is less than other popular morphological
biometric characteristics like face print, iris, and finger print recognition due to high rate
of intraclass variation or high Failure to Enroll Rate (FER). So, this technique demands
higher level of security. In this chapter, we are interested in investigating the integration
of the soft biometric features, gender and age group, with the existing keystroke dynamics
user authentication systems proposed by [1–3].
We have investigated the probability of predicting gender and age group based on typing
pattern. This is possible if all the patterns of certain peer are similar and dissimilar from
one peer to another. As per the research direction, it is possible to predict the gender with
88.55% to 95.04% accuracy based on typing pattern on the keyboard. Similarly, we also
obtained 84.75% accuracy with regard to gender based on typing pattern on touch screen.
If we use only gender information as an extra feature, then we can achieve 3.5% to 7.72%
of accuracy. Similarly, the age group (18–30/30+ years) can also be predicted based on
typing pattern. We obtained 86.87% to 94.68% accuracy with regard to prediction of age
group (18–30/30+ years) based on typing pattern on keyboard, whereas 84.75% of accuracy
was obtained by analyzing the typing pattern on touch screen with regard to age group
(7–29/30–65 years). We also analyzed the age group (≤18/18+ years) and obtained 89.2% to
92% accuracy based on typing pattern on touch screen.
These two biometric traits have low discriminating power but can be used as additional
soft biometric features to reduce the error rate in keystroke dynamics user authentication
systems. This technique can also be used in e-commerce sites to reach out to the right
client to avoid adverse products more efficiently based on the gender and age group.
There are many application areas where automatic gender and age group identification
methods can be applied like any surveillance system, online automatic user account
profiling, and protection of minors from online threats.
The main objective of this study is to develop a model that can identify the gender and
age group of users through the way of typing on a computer keyboard and touching a
computer screen for a predefined text, and it increases the accuracy by recognizing this
soft biometric information as additional features in keystroke dynamics user authentication systems.
The major goal and contributions of this chapter are as follows:
• Develop an efficient model to recognize the gender and age group automatically
from typing pattern.
• Validate the model on keystroke dynamics dataset collected through a touch
screen device.