State duration modelling in hidden Markov models

Vaseghi, Saeed

doi:10.1016/0165-1684(94)00088-h

Cited by 56 publications

(35 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, it also requires a considerable amount of memory space for the storage of all the duration distributions. On the other hand, the parametric method allows some specific probability density functions, such as Poisson (Russell & Moore, 1985;Russell & Cook, 1987), gamma (Levinson, 1986;Burshtein, 1995), Gaussian (Rabiner, 1989;Burshtein, 1995) and bounded density functions (Gu et al, 1991;Kim, Yoon & Youn, 1994;Vaseghi, 1995;Power, 1996;Laurila, 1997) to be used to model the state duration distributions explicitly, which requires only a few parameters to completely specify its distribution. It is inevitable that there are some drawbacks in the use of the parametric approach.…”

Section: Introductionmentioning

confidence: 99%

Improvement of noisy speech recognition using a proportional alignment decoding algorithm in the training phase

Hung

Wang

1998

Computer Speech & Language

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Improvement of noisy speech recognition using a proportional alignment decoding algorithm in the training phase

Hung

Wang

1998

Computer Speech & Language

View full text Add to dashboard Cite

“…Each activity will be placed in a HAM leaf node (sensor level) according to its day and time of occurrence, and will have a start time and event durations assigned to it. There have been earlier approaches to modeling durations of states in a Markov chain such as the approach by Vaseghi [14] in which state transition probabilities are conditioned on how long the current stated has been occupied. In our model, for each activity at each time node, we describe the start time and the duration of each individual event using a normal distribution that will be updated every time new data is collected.…”

Section: Temporal Informationmentioning

confidence: 99%

An Adaptive Sensor Mining Framework for Pervasive Computing Applications

Rashidi

Cook

2010

Knowledge Discovery From Sensor Data

View full text Add to dashboard Cite

Considering the wealth of sensor data in pervasive computing applications, mining sequences of sensor events brings unique challenges to the KDD community. The challenge is heightened when the underlying data source is dynamic and the patterns change. In this work, we introduce a new adaptive mining framework that detects patterns in sensor data, and more importantly, adapts to the changes in the underlying model. In our framework, the frequent and periodic patterns of data are first discovered by the Frequent and Periodic Pattern Miner (FPPM) algorithm; and then any changes in the discovered patterns over the lifetime of the system are discovered by the Pattern Adaptation Miner (PAM) algorithm, in order to adapt to the changing environment. This framework also captures vital context information present in pervasive computing applications, such as the startup triggers and temporal information. In this paper, we present a description of our mining framework and validate the approach using data collected in the CASAS smart home testbed. Categories and Subject Descriptors General TermsAlgorithms, Design, Experimentation, Human Factors. KeywordsSensor data mining, Sequential mining, Pervasive computing applications, Smart environments, Adaptation. INTODUCTIONWith remarkable recent progress in computing power, networking equipment, sensors, and various data mining methods, we are steadily moving towards ubiquitous and pervasive computing, into a world entangled with abundant sensors and actuators. As a result, there is a wealth of sensor data that can be analyzed with the goal of identifying interesting patterns. For example, by discovering repetitive sequences (frequent or periodic), modeling their temporal constraints and learning their expected utilities, we can intelligently automate a variety of tasks such as repetitive daily tasks in homes, or assembly sequences in manufacturing floors. Mining sequences of sensor events brings unique challenges to the KDD community, and the challenge is even heightened more when the underlying data source is dynamic and the patterns change.In this work, we introduce a new adaptive mining framework for use with pervasive computing applications, which will detect and adapt to changing patterns in the sensor data. These patterns can be expressed as a set of time ordered sequences, and discovering and adapting to the changes in such sequences can be achieved by using a sequence mining algorithm tailored to the special domain requirements of pervasive computing applications. The pervasive computing special requirements include utilizing context information such as startup triggers and temporal information, a unified framework for discovering periodic and frequent patterns, and most importantly adaptation over the lifetime of the system. Startup triggers are events that can trigger another action, e.g. a person entering a room can act as a potential startup trigger for the light in the room to be turned on. Triggers, which are absent in most traditional data mining methods, ...

show abstract

“…Another important source of temporal information is the start time and duration distributions of events of an activity. There have been earlier approaches to modeling durations of states in a Markov decision process such as the approach by Vaseghi [13] in which state transition probabilities are conditioned on how long the current stated has been occupied. In HAM, we model the start time of the first event of an activity and durations of all events of an activity using a normal distribution.…”

Section: Hierarchical Activity Model: Hammentioning

confidence: 99%

Keeping the intelligent environment resident in the loop

Rashidi

Cook

2008

4th International Conference on Intelligent Environments (IE 08)

View full text Add to dashboard Cite

Recent technological advancements have increased the likelihood that smart home technologies will become part of our everyday environments. However, many of these technologies are brittle and do not adapt to the user's wishes or changes in residents' habits and lifestyle. Here we introduce CASAS, an adaptive smart home that utilizes machine learning techniques to discover patterns in user behavior and to automatically mimic these patterns. Our goal is to keep the resident in control of the automation. Users can provide feedback on proposed automation activities, modify the automation policies, and introduce new requests. In addition, CASAS can discover changes in resident's behavior patterns automatically. In this paper we describe the CASAS technologies and evaluate its ability to adapt to the resident's activities and requests.

show abstract

State duration modelling in hidden Markov models

Cited by 56 publications

References 9 publications

Improvement of noisy speech recognition using a proportional alignment decoding algorithm in the training phase

Improvement of noisy speech recognition using a proportional alignment decoding algorithm in the training phase

An Adaptive Sensor Mining Framework for Pervasive Computing Applications

Keeping the intelligent environment resident in the loop

Contact Info

Product

Resources

About