Abstract
Two SOM ANN approaches were used in a study of Australian railway drivers (RDs) to classify RDs’ sleep/wake states and their sleep duration time series profiles over 14 days follow-up. The first approach was a feature-based SOM approach that clustered the most frequently occurring patterns of sleep. The second created RD networks of sleep/wake/duty/break feature parameter vectors of between-states transition probabilities via a multivariate extension of the mixture transition distribution (MTD) model, accommodating covariate interactions. SOM/ANN found 4 clusters of RDs whose sleep profiles differed significantly. Generalised Additive Models for Location, Scale and Shape of the 2 sleep outcomes confirmed that break and sleep onset times, break duration and hours to next duty are significant effects which operate differentially across the groups. Generally sleep increases for next duty onset between 10 am and 4 pm, and when hours since break onset exceeds 1 day. These 2 factors were significant factors determining current sleep, which have differential impacts across the clusters. Some drivers groups catch up sleep after the night shift, while others do so before the night shift. Sleep is governed by the RD’s anticipatory behaviour of next scheduled duty onset and hours since break onset, and driver experience, age and domestic scenario. This has clear health and safety implications for the rail industry.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
D. Darwent et al., Managing fatigue: it really is about sleep. Accid. Anal. Prev. 82, 20–26 (2015)
C. Bearman et al., Evaluation of Rail Technology: A Practical Human Factors Guide (Ashgate, United Kingdom, 2013)
D. Dawson, Fatigue research in 2011: from the bench to practice. Accid. Anal. Prev. 45(Suppl), 1–5 (2012)
J. Dorrian et al., Work hours, workload, sleep and fatigue in Australian Rail Industry employees. Appl Ergon 42, 202–209 (2011)
D.F. Dinges, An overview of sleepiness and accidents. J. Sleep Res. 4, 4–14 (1995)
S. Folkard, P. Tucker, Shift work, safety and productivity. Occup Med (Lond) 53, 95–101 (2003)
S. Folkard et al., Shiftwork: safety, sleepiness and sleep. Ind. Health 43, 20–23 (2005)
A. Kosmadopoulos et al., The effects of a split sleep-wake schedule on neurobehavioural performance and predictions of performance under conditions of forced desynchrony. Chronobiol. Int. 31, 1209–1217 (2014)
J.L. Paterson et al., Beyond working time: factors affecting sleep behaviour in rail safety workers. Accid. Anal. Prev. 45(Suppl), 32–35 (2012)
M. A. Short, et al., A systematic review of the sleep, sleepiness, and performance implications of limited wake shift work schedules. Scandinavian J. Work Environ. Health (2015)
G.D. Roach et al., The amount of sleep obtained by locomotive engineers: effects of break duration and time of break onset. Occup. Environ. Med. 60, e17 (2003)
D. Darwent et al., A model of shiftworker sleep/wake behaviour. Accid. Anal. Prev. 45(Suppl), 6–10 (2012)
S.W. Kim, Bayesian and non-Bayesian mixture paradigms for clustering multivariate data: time series synchrony tests. PhD, University of South Australia, Adelaide, Australia (2011)
P. Gander et al., Fatigue risk management: organizational factors at the regulatory and industry/company level. Accid. Anal. Prev. 43, 573–590 (2011)
F.M. Fischer et al., 21st International symposium on shiftwork and working time: the 24/7 society–from chronobiology to practical life. Chronobiol. Int. 31, 1093–1099 (2014)
R.A. Rigby, D.M. Stasinopoulos, Generalized additive models for location, scale and shape (with discussion). Appl. Stat. 54, 507–554 (2005)
A.A. Borbely, A two process model of sleep regulation. Hum Neurobiol 1, 195–204 (1982)
S. Daan et al., Timing of human sleep: recovery process gated by a circadian pacemaker. Am. J. Physiol. 246, R161–183 (1984)
G. Belenky et al., Patterns of performance degradation and restoration during sleep restriction and subsequent recovery: a sleep dose-response study. J. Sleep Res. 12, 1–12 (2003)
H.P. Van Dongen et al., The cumulative cost of additional wakefulness: dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep 26, 117–126 (2003)
A. A. Borbely, et al., Sleep initiation and initial sleep intensity: interactions of homeostatic and circadian mechanisms. J. Biol. Rhythms 4, 149–160 (1989)
P. Achermann et al., A model of human sleep homeostasis based on EEG slow-wave activity: quantitative comparison of data and simulations. Brain Res. Bull. 31, 97–113 (1993)
D.G. Beersma, P. Achermann, Changes of sleep EEG slow-wave activity in response to sleep manipulations: to what extent are they related to changes in REM sleep latency? J. Sleep Res. 4, 23–29 (1995)
M.E. Jewett, R.E. Kronauer, Refinement of a limit cycle oscillator model of the effects of light on the human circadian pacemaker. J. Theor. Biol. 192, 455–465 (1998)
R.E. Kronauer et al., Quantifying human circadian pacemaker response to brief, extended, and repeated light stimuli over the phototopic range. J. Biol. Rhythms 14, 500–515 (1999)
P. McCauley et al., A new mathematical model for the homeostatic effects of sleep loss on neurobehavioral performance. J. Theor. Biol. 256, 227–239 (2009)
T. Akerstedt, S. Folkard, Predicting duration of sleep from the three process model of regulation of alertness. Occup. Environ. Med. 53, 136–141 (1996)
S.R. Hursh, et al., Fatigue models for applied research in warfighting. Aviat. Space Environ. Med. 75, A44-53; discussion A54-60 (2004)
M. Moore-Ede et al., Circadian alertness simulator for fatigue risk assessment in transportation: application to reduce frequency and severity of truck accidents. Aviat. Space Environ. Med. 75, A107–118 (2004)
K.J. Kandelaars, et al., A review of bio-mathematical fatigue models: where to from here?, in 2005 International Conference on Fatigue Management in Transport Operations (2005), pp. 11–15
D. Darwent et al., Prediction of probabilistic sleep distributions following travel across multiple time zones. Sleep 33, 185–195 (2010)
A.A. Borbely, A two process model of sleep regulation. Hum. Neurobiol. 1, 195–204 (1982)
D. Darwent, The sleep of transportation workers in Australian rail and aviation operations. PhD, School of Education Arts and Social Sciences, University of South Australia, Adelaide (2006)
A.A. Borbely, et al., Sleep initiation and initial sleep intensity: interactions of homeostatic and circadian mechanisms. J. Biol. Rhythms. 4, 37–48 (1989)
S. Daan et al., Timing of human sleep: recovery process gated by a circadian pacemaker. Am. J. Physiol.: Regul. Integr. Comp Physiol. 246, 161–183 (1984)
N.J. Wesensten et al., Does sleep fragmentation impact recuperation?A review and reanalysis. J. Sleep Res. 8, 237–245 (1999)
D.F. Dinges et al., Cumulative sleepiness, mood disturbance, and psychomotor vigilance performance decrements during a week of sleep restricted to 4–5 h per night. Sleep 20, 267–277 (1997)
R.T. Wilkinson et al., Performance following a night of reduced sleep. Psychonomis Sci. 5, 471–472 (1966)
M. Lumley et al., The alerting effects of naps in sleep-deprived subjects. Psychophysiology 23, 403–408 (1986)
P. Achermann, A.A. Borbély, Mathematical models of sleep regulation. Front. Biosci.: J. Virtual Libr. 8, s683–s693 (2003)
D.G. Beersma, Models of human sleep regulation. Sleep Med. Rev. 2, 31–43 (1998)
D.J. Dijk, R.E. Kronauer, Commentary: models of sleep regulation: successes and continuing challenges. J. Biol. Rhythms 14, 569–573 (1999)
M.M. Mallis et al., Summary of the key features of seven biomathematical models of human fatigue and performance. Aviat. Space Environ. Med. 75, A4–A14 (2004)
M. Koslowsky, H. Babkoff, Meta-analysis of the relationship between total sleep deprivation and performance. Chronobiol. Int. 9, 132–136 (1992)
M.H. Bonnet, D.L. Arand, Level of arousal and the ability to maintain wakefulness. J. Sleep Res. 8, 247–254 (1999)
S.M. Doran et al., Sustained attention performance during sleep deprivation, evidence of state instability. Arch. Ital. Biol. 139, 253–267 (2001)
D.A. Drummond et al., Why highly expressed proteins evolve slowly. Proc. Natl. Acad. Sci. USA 102, 14338–14343 (2005)
X. Zhou et al., Sleep, wake and phase dependent changes in neurobehavioral function under forced desynchrony. Sleep 34, 931–941 (2011)
D.A. Cohen, et al., Uncovering residual effects of chronic sleep loss on human performance. Sci. Transl. Med. 2, 1413 (2010)
E.J. Silva et al., Circadian and wake-dependent influences on subjective sleepiness, cognitive throughput, and reaction time performance in older and young adults. Sleep 33, 481–490 (2010)
I. Philibert, Sleep loss and performance in residents and nonphysicians; a meta-analytic examination. Sleep 28, 1393–1402 (2005)
D.J. Kim et al., The effect of total sleep deprivation on cognitive functions in normal adult male subjects. Int. J. Neurosci. 109, 127–137 (2001)
Y. Harrison, J.A. Horne, The impact of sleep deprivation on decision making: a review. J. Exp. Psychol. Appl. 6, 236–249 (2000)
G. Belenky et al., Patterns of performance degradation and restoration during sleep restriction and subsequent recovery: a sleep dose-response study. J. Sleep Res. 12, 1–12 (2003)
M. Gillberg, T. Akerstedt, Sleep restriction and SWS-suppression: effects on daytime alertness and night-time recovery. J. Sleep Res. 3, 144–151 (1994)
L. Rosenthal et al., Level of sleepiness and total sleep time following various time in bed conditions. Sleep 16, 226–232 (1993)
H.P.A. Van Dongen et al., The cumulative cost of additional wakefulness: dose-response effects on neurobehavioral functions and sleep physiology from chronic sleep restriction and total sleep deprivation. Sleep 26, 117–126 (2003)
A. Hak, R. Kampmann, Working irregular hours: complaints and slate of fitness of railway personnel, ed. by A. Reinberg, et al. Night and Shiftwork: Biological and Social Aspects (Pergamon Press, Oxford, 1981), pp. 229–236
M. Härmä et al., The effect of an irregular shift system on sleepiness at work in train drivers and railway traffic controllers. J. Sleep Res. 11, 141–151 (2002)
L. Torsvall, T. Akerstedt, Sleepiness on the job: continuously measured EEG changes in train drivers. Electroencephalogr. Clin. Neurophysiol. 66, 502–511 (1987)
G.D. Edkins, C.M. Pollock, The influence of sustained attention on railway accidents. Accid. Anal. Prev.: Spec. Issue, Fatigue Transp. 29, 533–539 (1997)
K. Kogi, Sleep problems in night and shift work. J. Hum. Ergol. (Tokyo) 11(suppl), 217–231 (1982)
P. Naitoh et al., Health effects of sleep deprivation. Occup. Med. 5, 209–237 (1990)
J. Rutenfranz, Occupational health measures for night- and shiftworkers. J. Hum. Ergol. (Tokyo) 11(suppl), 67–86 (1982)
A.J. Scott, J. LaDou, Shiftwork: effects on sleep and health with recommendations for medical surveillance and screening. Occup. Med. 5, 273–299 (1990)
N. Kurumatani et al., The effects of frequently rotating shiftwork on sleep and the family life of hospital nurses. Ergonomics 37, 995–1007 (1994)
P. Knauth, J. Rutenfranz, in Duration of sleep related to the type of shiftwork, ed. by A. Reinberg, et al. Advances in the Biosciences, Vol 30. Night and shiftwork: Biological and Social Aspects (Pergamo Press, New York, 1980), pp. 161–168
R.R. Mackie, J.C. Miller, Effects of hours of service regularity of schedules and cargo loading on truck and bus driver fatigue, US Department of Transport, vol. HS-803 (1978), p. 799
C.D. Wylie, et al., Commercial Motor Vehicle Driver Rest Periods and Recovery of Performance (Transportation Development Centre, Safety and Security, Transport Canada, Montreal, Quebec, Canada, 1997)
P. Knauth, J. Rutenfranz, The effects of noise on the sleep of night-workers, ed. by P. Colquhoun, et al. Experimental Studies of Shift Work (Westedeutscher Verlag, Opladen, 1975), pp. 57–65
K.R. Parkes, Sleep patterns, shiftwork, and individual differences: a comparison of onshore and offshore control-room operators. Ergonomics 37, 827–844 (1994)
T. Akerstedt, Shift work and disturbed sleep/wakefulness. Soc. Occup. Med. 53, 89–94 (2003)
T. Akerstedt et al., Spectral analysis of sleep electroencephalography in rotating three-shift work. Scand. J. Work Environ. Health 17, 330–336 (1991)
G. Costa, The impact of shift and night work on health. Appl. Ergon. 27, 9–16 (1996)
G. Costa, The problem: shiftwork. Chronobiol. Int. 14, 89–98 (1997)
C. Cruz et al., Clockwise and counterclockwise rotating shifts: effects on vigilance and performance. Aviat. Space Environ. Med. 74, 606–614 (2003)
C. Cruz et al., Clockwise and counterclockwise rotating shifts: effects on sleep duration, timing, and quality. Aviat. Space Environ. Med. 74, 597–605 (2003)
F.M. Fischer et al., Day- and shiftworkers’ leisure time. Ergonomics 36, 43–49 (1993)
M. Frese, C. Harwich, Shiftwork and the length and quality of sleep. J. Occup. Environ. Med. 26, 561–566 (1984)
V.H. Goh et al., Circadian disturbances after night-shift work onboard a naval ship. Mil. Med. 165, 101–105 (2000)
R. Loudoun, P. Bohle, Work/non-work conflict and health in shiftwork: relationships with family status and social support. Int. J. Occup. Environ. Health 3(suppl 2), 71–77 (1997)
J. Walker, Social problems of shift work, ed. by S. Folkard and T. Monk, in Hours of Work—Temporal Factors in Work Scheduling (Wiley, New York, 1985), pp. 211–225
M.J. Blake, Relationship between circadian rhythm of body temperature and introversion-extraversion. Nature 215, 896–897 (1967)
W.P. Colquhoun, Biological Rhythms and Human Performance (Academic, London, 1971)
G.S. Richardson et al., Circadian variation of sleep tendency in elderly and young adult subjects. Sleep 5(suppl 2), S82–S94 (1982)
W.B. Webb, D.F. Dinges, Cultural perspective on napping and the siesta, ed. by D.F. Dinges, R.J. Broughton, in Sleep and Alertness: Chronobiological, Behavioral, and Medical Aspects of Napping (Raven Press, New York, 1989), pp. 247–265
E. Hoddes et al., The development and use of the stanford sleepiness scale. Psychophysiology 9, 150 (1972)
H. Nabi, et al., Awareness of driving while sleepy and road traffic accidents: prospective study in GAZEL cohort. BMJ 333, 75 (2006)
L.A. Reyner, J.A. Horne, Falling asleep whilst driving: are drivers aware of prior sleepiness? Int. J. Legal Med. 111, 120–123 (1998)
S.W. Kim, et al., Modelling the flowering of four Eucalypts species via MTDg with interactions, in 18th World IMACS Congress and International Congress on Modelling and Simulation, Cairns, Australia, 2009, pp. 2625–2631
I. Hudson, et al., Modelling the flowering of four eucalypt species using new mixture transition distribution models, ed. by I.L. Hudson, M.R. Keatley in Phenological Research (Springer, Netherlands, 2010), pp. 299–320
I.L. Hudson et al., Using self-organising maps (SOMs) to assess synchronies: an application to historical eucalypt flowering records. Int. J. Bio-meteorol. 55, 879–904 (2011)
I.L. Hudson, et al., Modelling lagged dependency of flowering on current and past climate on Eucalypt flowering: a mixture transition state approach, in International Congress of Biometeorology, Auckland, New Zealand, 2011, pp. 239–244
A. Berchtold, A.E. Raftery, The mixture transition distribution model for high- order Markov chains and non-Gaussian time series. Stat. Sci. 17, 328–356 (2002)
B.H. Junker et al., VANTED: a system for advanced data analysis and visualization in the context of biological networks. BMC Bioinformatics 7, 109 (2006)
I.L. Hudson, et al., SOM clustering and modelling of Australian railway drivers’ sleep, wake, duty profiles, in 28th International Workshop on Statistical Modelling, Palermo, Italy, 2013, pp. 177–182
I.L. Hudson, et al., Climate effects and thresholds for flowering of eight Eucalypts: a GAMLSS ZIP approach, in 19th International Congress on Modelling and Simulation, Perth, Australia, 2011, pp. 2647–2653
I. Hudson, et al., Climatic Influences on the flowering phenology of Four Eucalypts: a GAMLSS approach, in ed. by I.L. Hudson, M.R. Keatley, in Phenological Research (Springer Netherlands, 2010), pp. 209–228
T. Kohonen, Self-organizing maps, 3rd edn. (Springer, Berlin, 2001)
P.N. Nguyen et al., Living standards of Vietnamese provinces: a Kohonen map. Case Stud. Bus. Ind. Govern. Stat. 22, 109–113 (2009)
J.A. Sleep, I.L. Hudson, Comparison of self-organising maps, mixture, K-means, and hybrid approaches to risk classification of passive railway crossings, in 23rd International Workshop on Statistical Modelling, Utrecht, Netherlands, 2008
C. Fraley, et al., MCLUST Version 4 for R: normal mixture modeling and model-based clustering, classification, and density estimation, Department of Statistics, University of Washington Technical Report No. 504 (2013)
J.F. Roddick, M. Spiliopoulou, A survey of temporal knowledge discovery paradigms and methods. IEEE Trans. Knowl. Data Eng. 14, 750–767 (2002)
P. Cheeseman and J. Stutz, Bayesian classification (AutoClass): theory and results, U.M. Fayyard, et al., in Advances in Knowledge Discovery and Data Mining (AAAI/MIT Press, Cambridge, 1996)
C. Biernacki et al., Assessing a mixture model for clustering with the integrated completed likelihood. IEEE Trans. Pattern Anal. Mach. Intell. 22, 719–725 (2000)
D.B. Reusch, et al., North Atlantic climate variability from a self-organizing map perspective. J. Geophys. Res. 112 (2007)
A.F. Costa, Clustering and visualizing SOM results, in 11th International Conference on Intelligent Data Engineering and Automated Learning—IDEAL 2010. Lecture Notes in Computer Science. vol. 6283 (Springer, Berlin, 2010), pp. 334–343
J. Vesanto, E. Alhoniemi, Clustering of the self-organizing map. IEEE Trans. Neural Netw. 11, 586–600 (2000)
H. Yin, Learning nonlinear principal manifolds by self-organising maps, ed. by A. Gorban, et al. in Principal Manifolds for Data Visualization and Dimension Reduction, vol. 58 (Springer, Berlin, 2008), pp. 68–95
J.C. Fort, SOM’s mathematics. Neural Netw. 19, 812–816 (2006)
C. Klukas, The VANTED software system for transcriptomics, proteomics and metabolomics analysis. J. Pestic Sci. 31, 289–292 (2006)
D.M. Stasinopoulos, R.A. Rigby, Generalized additive models for location scale and shape (GAMLSS) in R. J. Stat. Softw. 23, 1–46 (2007)
R.A. Rigby, D.M. Stasinopoulos, MADAM macros to fit mean and dispersion additive models, ed. by A. Scallan, G. Morgan, in A. Scallan, G. Morgan, GLIM4 Macro Library Manual, Release 2.0 (Numerical Algorithms Group, Oxford, 1996), pp. 68–84
R.A. Rigby, D.M. Stasinopoulos, Mean and dispersion additive models, ed. by W. Hardle, M.G. Schimek, in Statistical Theory and Computational Aspects of Smoothing (Physica-Verlag, Heidelberg, 1996), pp. 215–230
A. Berchtold, Markov chain computation for homogeneous and non-homogeneous data: MARCH 1.1 Users Guide. J. Stat. Softw. (2001)
A. Berchtold, March v.3.00. Markovian models Computation and Analysis Users guide (2006). http://www.andreberchtold.com/march.html
A. Berchtold, March v.2.01. Markovian models Computation and Analysis Users guide (2004). http://www.andreberchtold.com/march.html
S.W. Kim, et al., Modelling and synchronization of four Eucalypt species via Mixed Transition Distribution (MTD) and Extended Kalman Filter (EKF), in 23rd International Workshop on Statistical Modelling, Utrecht, Netherlands, 2008, pp. 287–292
D.A. Fournier, AD Model Builder, Version 5.0.1. (Otter Research Ltd., Canada, 2000)
D.A. Fournier, et al., AD Model Builder: using automatic differentiation for statistical inference of highly parameterized complex nonlinear models. Optim. Methods Softw. 27, 233–249 (2012)
I.L. Hudson, et al., Climatic influences on the flowering phenology of four Eucalypts: a GAMLSS approach, in 18th World IMACS Congress and International Congress on Modelling and Simulation. Modelling and Simulation Society of Australia and New Zealand, Cairns, Australia, 2009, pp. 2611–2617
I.L. Hudson, et al., Climate impacts on sudden infant death syndrome: a GAMLSS approach, in 23rd International Workshop on Statistical Modelling (IWSM), Utrecht, Netherlands, 2008, pp. 277–280
T. Akerstedt et al., Accounting for partial sleep deprivation and cumulative sleepiness in the Three-Process Model of alertness regulation. Chronobiol. Int. 25, 309–319 (2008)
P.M. Krueger, E.M. Friedman, Sleep duration in the United States: a cross-sectional population-based study. Am. J. Epidemiol. 169, 1052–1063 (2009)
D. Nur, et al., Multivariate Gaussian hidden Markov models for sleep profiles of railway drivers, in Australian Statistical Conference, Adelaide, Australia, July 2012, p. 159
A.A. Markov, Extension of the limit theorems of probability theory to a sum of variables connected in a chain (reprinted in Appendix B), ed. by R. Howard, in Dynamic Probabilistic Systems volume 1: Markov Chains (Wiley, New York, 1971)
N. Goldman, Z. Yang, A codon-based model of nucleotide substitution for protein-coding DNA sequences. Mol. Biol. Evol. 11, 725–736 (1994)
L. Muscariello et al., Markov models of internet traffic and a new hierarchical MMPP model. Comput. Commun. 28, 1835–1851 (2005)
S.F. Gray, Modeling the conditional distribution of interest rates as a regime-switching process. J. Finan. Econ. 42, 27–62 (1996)
C.J. Spanos, et al., The economic impact of choosing off-line, inline or in situ metrology deployment in semiconductor manufacturing, in Semiconductor Manufacturing Symposium, 2001 IEEE International, 2001, pp. 37–40
W.A.V. Clark, Markov chain analysis in geography: an application to the movement of rental housing areas. Ann. Assoc. Am. Geogr. 55, 351–359 (1965)
B. Mau et al., Bayesian phylogenetic inference via Markov Chain Monte Carlo methods. Biometrics 55, 1–12 (1999)
N.G. van Kampen, Stochastic Processes in Physics and Chemistry, 3rd edn. (Elsevier, Amsterdam, 2007)
J. Claerbout, Fundamentals of Geophysical Data Processing: with Applications to Petroleum Prospecting (McGraw-Hill Inc., New York, 1976)
A. Berchtold, G. Sackett, Markovian models for the developmental study of social behavior. Am. J. Primatol. 58, 149–167 (2002)
D. Draper, Inference and hierarchical modeling in the social sciences. J. Educ. Behav. Stat. 20, 115–147 (1995)
K. Verbeurgt, et al., Extracting patterns in music for composition via Markov chains, in 17th International Conference On Innovations in Applied Artificial Intelligence, Ottawa, Canada, 2004, pp. 1123–1132
P. Bremaud, Markov Chains: Gibbs Fields, Monte Carlo Simulation, and Queues (Springer, New York, 1999)
E. Seneta, Markov and the birth of chain dependence theory. Int. Stat. Rev. 64, 255–263 (1996)
A.E. Raftery, A model for high-order Markov chains. J. R. Stat. Soc. Ser. B (Stat. Methodol.) 47, 528–539 (1985)
A. Berchtold, Autoregressive modelling of Markov chains, in 10th International Workshop on Statistical Modelling (IWSM) (Springer, New York, 1995), pp. 19–26
A. Berchtold, Swiss health insurance system: mobility and costs. Health Syst. Sci. 1, 291–306 (1997)
A. Berchtold, The double chain Markov model. Commun. Stat.: Theory Methods 28, 2569–2589 (1999)
A. Berchtold, Estimation in the mixture transition distribution model. J. Time Ser. Anal. 22, 379–397 (2001)
A. Berchtold, A.E. Raftery, The mixture transition distribution model for high- order markov chains and non-gaussian time series. Stat. Sci. 17, 328–356 (2002)
A. Berchtold, High-order extensions of the double chain Markov model. Stoch. Models 18, 193–227 (2002)
A. Berchtold, Mixture transition distribution (MTD) modeling of heteroscedastic time series. Comput. Stat. Data Anal. 41, 399–411 (2003)
A. Berchtold. (2004, March v.2.01. Markovian models computation and analysis users guide. http://www.andreberchtold.com/march.html
A. Berchtold. (2006, March v.3.00. Markovian models computation and analysis users guide. http://www.andreberchtold.com/march.html
K. Fokianos, B. Kedem, Regression theory for categorical time series. Stat. Sci. 18, 357–376 (2003)
S. Lèbre, P.-Y. Bourguignon, An EM algorithm for estimation in the mixture transition distribution model. J. Stat. Comput. Simul. 78, 713–729 (2008)
J. Luo, H.-B. Qiu, Parameter estimation of the WMTD model. Appl. Math.: J. Chin. Univ. 24, 379–388 (2009)
C.S. Wong, W.K. Li, On a mixture autoregressive model. J. R. Stat. Soc. Ser. B (Stat. Methodol.) 62, 95–115 (2000)
S.W. Kim, et al., Modelling and synchronization of four Eucalypt species via Mixed Transition Distribution (MTD) and Extended Kalman Filter (EKF), in 23rd International Workshop on Statistical Modelling (IWSM), Utrecht, Netherlands, 2008, pp. 287–292
S.W. Kim, Bayesian and non-Bayesian mixture paradigms for clustering multivariate data: time series synchrony tests. PhD, University of South Australia, Adelaide, Australia (2011)
I. Hudson, et al., Modelling the flowering of four Eucalypt species using new mixture transition distribution models, ed. by I.L. Hudson, M.R. Keatley, in Phenological Research (Springer, Netherlands, 2010), pp. 299–320
S.W. Kim, et al., MTD analysis of flowering and climatic states, in 20th International Workshop on Statistical Modelling, Sydney, Australia, 2005, pp. 305–312
I.L. Hudson, et al., Modelling the flowering of four Eucalypt eucalypts species using new mixture transition distribution models, ed. by I.L. Hudson, M.R. Keatley, in Phenological Research: Methods for Environmental and Climate Change Analysis (Springer, Dordrecht, 2010), pp. 315–340
S.W. Kim, et al., Analysis of sleep/wake and duty profiles of railway drivers using GAMLSS with interactions, in Australian Statistical Conference, Adelaide, South Australia, July 2012, p. 121
S.W. Kim, et al., Mixture transition distribution analysis of flowering and climatic states, in 20th International Workshop on Statistical Modelling, Sydney, Australia, 2005, pp. 305–312
S.W. Kim, et al., Modelling the flowering of four eucalypts species via MTDg with interactions, in 18th World International Association for Mathematics and Computers in Simulation (IMACS) Congress and International Congress on Modelling and Simulation. Modelling and Simulation, Society of Australia and New Zealand and, Cairns, Australia, 2009, pp. 2625–2631
S.W. Kim, et al., Modelling and synchronization of four Eucalypt species via Mixed Transition Distribution (MTD) and Extended Kalman Filter (EKF), in 23rd International Workshop on Statistical Modelling, Utrecht, Netherlands, 2008, pp. 287–292
D.A. Fournier, AD Model Builder, Version 5.0.1. (Otter Research Ltd., Canada, 2000)
D.A. Fournier, et al., AD Model Builder: using automatic differentiation for statistical inference of highly parameterized complex nonlinear models. Optim. Methods Softw. 27, 233–249 (2012)
N. Kilari et al., Block replacement modeling for electronic systems with higher order Markov chains. IUP J. Comput. Math. 4, 49–63 (2011)
N. Kilari et al., Reengineering treatment in block replacement decisions using higher order Markov chains. IUP J. Oper. Manag. 10, 22 (2011)
D.A. Fournier, AUTODIFF, A C++ Array Language Extension with Automatic Differentiation for use in Nonlinear Modeling and Statistics (Otter Research Ltd., Canada, 1996)
I.L. Hudson, et al., Climatic influences on the flowering phenology of four Eucalypts: a GAMLSS approach, ed. I.L. Hudson, M.R. Keatley, in Phenological Research: Methods for Environmental and Climate Change Analysis (Springer, Dordrecht, 2010), pp. 213–237
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix A: Mathematics of M-MTDg
Appendix A: Mathematics of M-MTDg
1.1 A.1 The Mixture Transition Distribution (MTD) Model
The classic Markov chain is a probabilistic model that represents dependences between successive observations of a random variable (usually over time). In this chapter a discrete state random variable (or multivariate analogues of) taking values in the finite set {1,…, m} is considered in the MTD formulation, which allows for a covariate interaction and modelling of high-order Markov chains (from a time series viewpoint). Markov chains are traditionally used to predict the current value as a function of the previous observations of this same variable (the so-called lagged dependency). The Markov chain was introduced by Andrej A. Markov [126] at the beginning of the twentieth century and has wide applicability in many areas such as: mathematical biology [127], internet applications [128], economics [129], meteorology [130], geography [131], biology [132], chemistry [133], physics [134], behavioural science [135], social sciences [136] and music [137]. For a comprehensive treatment of Markov chains and early applications see Bremaud [138]. Seneta [139] provides an account of Markov’s motivations including an excellent discussion of the early development of the theory.
Raftery [140] introduced the mixture transition distribution (MTD) model to model high-order Markov chains. Berchtold [135, 141–149] subsequently developed software (called Markovian Models Computation and Analysis, MARCH) to model Markov chains using a suite of methods including the MTD and the double chain Markov model. The MTD model has been applied to genomic sequence and time series data [147, 150–153].
The aim of part A of this chapter is to study the multivariate relationship between the probability of sleep with 4-states each of sleep onset times, break onset times, next duty onset times, break duration, and time of sleep via a multivariate mixture transition distribution (M-MTD) which accommodates a different transition matrix from each lag to the present (MTDg) analysis [154]. The issue of accommodating for interaction terms between covariates itself had not, till the work of Kim [155] and of Hudson et al. [156], been addressed in the MTD [140], MTDg nor MARCH [149] literature. Kim et al. [154, 157] and Hudson et al. [158] first introduced the concept of interactions based on work in this current chapter, along with GAMLSS [159].
The idea of the original mixture transition distribution model was to consider independently the effect of each lag to the present instead of considering the effect of the combination of lags as in pure Markov chain processes. The assumption behind the MTD model, namely the assumed equality of the transition matrices among different lags, is a strong assumption. We further extend the MTDg model to allow for interactions (between break, duration and next onset times with sleep times) to account for changes in the transition matrices amongst the differing covariates.
This work extends both the MARCH MTD software of Berchtold [149] and the previous work Hudson et al. [156] and of Kim et al. [160–162]. Our model is different to MARCH in terms of incorporating interactions between the covariates and also in its minimisation process [155]. It uses the AD Model BuilderTM [163], [164]. This M-MTD adaptation also utilises auto-differentiation as a minimisation tool, and was shown to be computationally less intensive than MARCH (see the papers of Kim et al. [161, 162] and Kim [155]). The AD Model BuilderTM platform (see Fournier et al. [164]) has great application in fisheries research and recently in computational mathematics and operations management (e.g. electronic systems models [165, 166]).
1.1.1 A.1.1 The MTD Model
Let {Y i } be a sequence of random variables taking values in the finite set N = {1,…, m}. In a lth-order Markov chain, the probability that i t , …, i o ∈ N depends on the combination of values taken by X t-l , …, X t-1 . In the MTD model, the contributions of the different lags are combined additively, as follows:
where i t , …, i o ∈ N, and where the probabilities \( q_{{i_{g} i_{0} }} \) are elements of a m × m transition matrix \( Q = \left[ {q_{{i_{g} i_{0} }} } \right] \), each row of which is a probability distribution (i.e., each row sums to 1 and the elements are nonnegative) and \( \lambda = \left( {\lambda_{l} , \ldots ,\lambda_{1} } \right)^{\prime } \) is a vector of lag parameters, such that
The vector λ is made subject to the following constraints, \( \sum\limits_{g = 1}^{l} {\lambda_{g} = 1} , \) and \( \lambda_{g} \ge 0 \). Equation (A.1) gives the probability for each individual combination of i l , …, i 0 . The model can also be written in matrix form, giving the whole distribution of X t [19].
Each row of the transition matrix Q is a probability distribution and as such sums to 1, where the matrix has m(m-1) independent parameters. In addition, a lth-order model has l lag parameters λ1,…, λ l , but only (l− 1) of them are independent. Thus a lth order MTD model has m (m – 1) + (l – 1) independent parameters, which is far more parsimonious than the corresponding fully parameterised Markov chain which has m l(m – 1) parameters. Moreover, each additional lag in a MTD model adds only one extra parameter. For the basic MTD model of Raftery [15], the same transition matrix Q is used to model the relationship between any of the lags and the present state.
1.1.2 A.1.2 The MTDg Model
Let {Y i } be a sequence of random variables taking values in the finite set N = {1,…, m}. In an lth-order Markov chain, the probability that \( X_{t} = i_{0} \,,\,i_{0} \in N \), depends on the combination of values taken by \( Y_{t - l} , \ldots ,Y_{t - 1} \). In the basic MTD model, the same transition matrix Q is used to model the relation between any of the lags and the present. The idea of the mixture transition distribution (MTD) model is to consider independently the effect of each lag to the present instead of considering the effect of the combination of lags (Fig. A.1), as in the case of the more traditional pure Markov chain process. The constraints imposed by the use of only one transition matrix to represent the relation between each lag and the present is sometimes too strong to allow good modeling of the real high-order transition matrix. In this case, it is possible to replace the basic MTD model by an MTDg model. The principle of the MTDg model is to use a different transition matrix of size (k × k) to represent the relationship between each lag and the present. The high-order transition probabilities are then written as follows,
where \( q_{{gi_{g} i_{0} }} \) is the transition probability from modality i g observed at time t-g and modality i 0 observed at time t in the transition matrix Q g associated with the gth lag. In addition to the lag weight vector \( \left[ {\lambda_{1} , \ldots ,\lambda_{f} } \right] \), the MTDg model implies the estimation of f transition matrices Q 1 ,…,Q f , for a total of f k (k − 1) + (f − 1) independent parameters. This is much more than was involved in the basic MTD model, but this number of parameters remains small compared to the number of independent parameters of a real fully parameterised fth order Markov chain; thus the MTDg and its extensions model prove useful in many situations, as shown in this chapter. Here, the contribution of each lag upon the present is considered independently. The MTD model thus approximates high-order Markov chains with far fewer parameters than the fully parameterised model. Though Markov chains are well suited to represent high-order dependencies between successive observations (of a random variable), as the order l of the chain and the number of possible values m increase, the number of independent parameters increases exponentially. The problem then becomes too large to be estimated efficiently, as is often the case for data sets of the size typically encountered in practice [145].
1.1.3 A.1.3 The MTDg Model with Interactions
The MTDg model with interactions can also have a different transition matrix of size (k × k) to represent the relationship between each lag and the present [155]. The high-order transition probabilities are then computed as follows
where \( \lambda_{f + e + u} \) is the weight for the interaction term, \( q_{{gi_{g} i_{0} }} \) is the transition probability from modality i g observed at time t–g and modality i 0 observed at time t in the transition matrix Q g associated with the gth lag, \( s_{{uv_{u} i_{0} }} \) is transition probability between covariate h 1 and covariate h 2 interaction term (\( v_{u} = d_{{h_{1} j_{{h_{1} }} }} \times d_{{h_{2} j_{{h_{2} }} }} \)) and Y t , and where \( \sum\limits_{g = 1}^{f + e + l} {\lambda_{g} = 1} \) and \( \lambda_{g} \ge 0 \).
1.1.4 A.1.4 Parameter Estimation
The parameters λ and q of the MTDg model can be estimated by minimising the negative the log-likelihood (NLL) of the model:
where \( n_{{i_{l} , \ldots ,i_{0} }} \) is the number of sequences of the form
To ensure that the model defines a high order Markov chain, the negative log-likelihood is minimised with respect to the constraints delineated above. ADMBTM was used to minimise the negative the log-likelihood (NLL). This uses auto-differentiation (AUTODIFF) [164] as the minimisation tool, shown to be computationally less intensive than MARCH [155]. Estimation algorithms relevant to this procedure can be found in Fournier [163, 167]. A major advantage of the new model is that its run-time is considerably shorter (less than one minute as compared with two days for the original MTD model [157]) and it can be run from a batch file in DOS. Hence, multiple models can be tested consecutively in remote mode. Outputs can also be appended into one file to allow easy access by any graphical software package [155].
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Hudson, I.L., Leemaqz, S.Y., Kim, S.W., Darwent, D., Roach, G., Dawson, D. (2016). SOM Clustering and Modelling of Australian Railway Drivers’ Sleep, Wake, Duty Profiles. In: Shanmuganathan, S., Samarasinghe, S. (eds) Artificial Neural Network Modelling. Studies in Computational Intelligence, vol 628. Springer, Cham. https://doi.org/10.1007/978-3-319-28495-8_11
Download citation
DOI: https://doi.org/10.1007/978-3-319-28495-8_11
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-28493-4
Online ISBN: 978-3-319-28495-8
eBook Packages: EngineeringEngineering (R0)