RWS (Random Walk Splitting): A Random Walk Based Discretization of Continuous Attributes

Hanaoka, Masaaki; Kobayashi, Masaki; Yamazaki, Haruaki

doi:10.1007/3-540-44533-1_18

Cited by 2 publications

(5 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 1. The value of characteristic parameters extracted from digitized PSG chart data * The value was set so that 700 to 800 instances were extracted in RWS [20], for which the ratio of correct answer is the highest. †…”

Section: Resultsmentioning

confidence: 99%

“…The classification accuracy is higher by approximately 2% in C4.5 with the pruning function than in ID3 and C45-NP (C4.5 without the pruning function) [34], by approximately 2% in RWS compared to ChiMerge [35] and MDLPC [36], which are typical discretization methods [20], and by approximately 3% in committee learning by bagging compared to the method using a single classifier [31].…”

Section: Discussionmentioning

confidence: 99%

“…In RWS [20], the splitting point is better if the difference between the two empirical distributions is maximized. The termination condition for the separation procedure is reduced to statistical testing based on the random-walk model [32].…”

Section: Rws: Discretization Of Continuous Attributes Based On Randommentioning

confidence: 99%

“…It is necessary in sleep stage scoring to discretize a large amount of continuous-valued attribute data (in the case of an epoch, which is for 20 seconds, there are 1440 instances for a sleep duration of 8 hours). Consequently, the continuous attributes are discretized using RWS [20], which is suited to such a case.…”

Section: Decision-tree Learningmentioning

confidence: 99%

“…By RWS discretization method [20] based on the random-walk model, the variables are converted from continuous attributes to discrete attributes. Then, based on the training instances extracted by sampling by the bootstrap method [21], a set of multiple small decision trees (a committee) is formed.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Automatic sleep stage scoring based on waveform recognition method and decision‐tree learning

Hanaoka

Kobayashi

Yamazaki

2002

Systems & Computers in Japan

Self Cite

View full text Add to dashboard Cite

SUMMARYThere have been many trials in which the waveform recognition method, which is intended to replace expert observation by computer processing, has been used to extract the features of biological signals during sleep, and to automatically score sleep stages based on feature parameters. This paper proposes a waveform recognition method which extracts the feature parameters based on the characteristics of the biological signal during sleep, and a method of automatic sleep stage scoring by decision-tree learning, which is currently considered to be one of the most successful machine learning methods in practice. As the first step in the method, the features corresponding to the state of the EEG, the EOG, and the EMG during sleep are compared to the features of characteristic waves such as the α-wave, δ-wave, sleep spindle, K-complex, and REM, and the feature parameters needed in order to judge the sleep stage are extracted. Using canonical discriminant analysis and discretization method RWS based on the random walk, the feature parameters are converted to a small number of discrete variables. Based on training instances, obtained by the bootstrap method, a set of multiple small decision trees (a committee) is formed, and the sleep stage is scored by majority decision in the classification results. The method is applied to the PSG chart digital data provided by the Japan Sleep Society, and the performance of the system is evaluated experimentally. It is verified that the proposed method is promising as a method of automatic sleep stage scoring with high accuracy, requiring little expenditure of time in learning and classification.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%