Skip to main content

Table 1 Available KCA algorithms based on free-text keystroke dynamics

From: TKCA: a timely keystroke-based continuous user authentication with short keystroke sequence in uncontrolled settings

Study

Feature

Method

Dataset

Performance

Length

(Monrose and Rubin 1997)

hold time, latency time

k-NN, Euclidean distance

collect

ACC: 23%

31

(Dowland and Furnell 2004)

n-graph flight time (n=2, 3),

mean, standard deviation

collect

FAR: 4.9%,

 
 

keyword-based keystroke latency

  

FRR: 0%,

 
    

ANIA: 6,390,

 
    

ANGA: 68,755

 

(Gunetti and Picardi 2005)

n-graph flight time (n=2, 3, 4)

‘R’ and ‘A’ Distance

collect

FAR: 0.005%

700-900

    

FRR: 4.833%

 

(Montalvão Filho and Freire 2006)

digraph flight time

Markov Chain

collect

EER: 12.7%

about 550

(Janakiraman and Sim 2007),

held time, inter-key time

Bhattacharyya Distance,

collect

ACC: 74-100%

1,500-100,000

(Sim and Janakiraman 2007)

 

Naive Bayes

   

(Davoudi and Kabir 2009;2010)

digraph flight time

HDE, ‘R’ and ‘A’ Distance

(Gunetti and Picardi 2005)

FAR: 0.07%, FRR: 15.2%

700-900

(Harun et al. 2010)

 

ANN, some Distance

(Montalvão Filho and Freire 2006)

EER: 22.9%

about 550

(Stewart et al. 2011)

hold time, latency time,

k-NN

collect

EER: 0.55%

about 6,000

 

percent usage of keys

    

(Al Solami et al. 2011)

most typed digraph

Clustering

(Gunetti and Picardi 2005)

ACC: 100%

700-900

(Messerman et al. 2011)

n-graph flight time (n=2, 3, 4)

‘R’ and ‘A’ Distance

collect

eFAR: 2.61%, iFAR: 2.02%, FRR: 1.84%

150-450

(Rahman et al. 2011)

latency time

‘R’ and ‘A’ Distance

collect

EER: 10-15%

854-1,836

(Ferreira and Santos 2012)

hold time, latency time,

mean, standard deviation

collect

EER of 1.4%

250

 

n-graph flight time (n=2, 3, 4)

    

(Bours 2012)

hold time, latency time

mean, standard deviation

collect

ANIA: 182

 

(Monaco et al. 2013)

hold time, latency time

mean, standard deviation

collect

EER: 3.7%

755

(Deutschmann et al. 2013)

 

Bayesian network

collect

ANIA: 760-950

 

(Ahmed and Traore 2013)

hold time, latency time,

ANN

collect

EER: 2.13%

500

 

n-graph flight time (n=2, 3)

    

(Kang and Cho 2015)

digraph flight time

12 different techniques

collect

EER: 5.64%

1,000

(Çeker and Upadhyaya 2016)

hold time, digraph flight time

One-class SVM

(Vural et al. 2014)

EER: 0

1,319-3,454

(Mondal and Bours 2017)

hold time, total time, latency time,

ANN, CPANN, SVM

collect

ANIA: 167 or larger

 
 

flight time, up-up time for digraph

    

(Alshehri et al. 2017;2018)

hold time, digraph flight time

DTW

collect, (Vural et al. 2014)

ACC: 98.39%, 97.32%

100

(Huang et al. 2017)

digraph flight time

KDE

(Gunetti and Picardi 2005),

EER: 3.48%, 3.36%, 1.95%, 7.59%

1,000

   

(Vural et al. 2014),

  
   

(Sun et al. 2016),

  
   

(Murphy et al. 2017)

  

(Ayotte et al. 2019)

digraph flight time

KDE, Energy Distance,

(Murphy et al. 2017)

EER: 35.1%, 15.3%, 6.3%, 3.6%

100, 200, 500, 1,000

  

Kolmogorov-Smirnov

   

(Xiaofeng et al. 2019)

keycode, hold time, latency

CNN, RNN

(Sun et al. 2016)

EER: 3.04

30

 

time, digraph flight time

   Â