Is the DTW “distance” really a metric? An algorithm reducing the number of DTW comparisons in isolated word recognition

Ruiz, Enrique Vidal; Nolla, Francisco Casacuberta; Segovia, Héctor Rulot

doi:10.1016/0167-6393(85)90058-5

Cited by 34 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is also worth noting that DTW does not generate a distance metric per se, as DTW violates the triangle inequality in some cases. Nevertheless, there is robust literature demonstrating that DTW distance is an appropriate distance metric in the overwhelming majority of cases (e.g [ 17 , 18 ].).…”

Section: Discussionmentioning

confidence: 99%

Female mice exhibit less overall variance, with a higher proportion of structured variance, than males at multiple timescales of continuous body temperature and locomotive activity records

Smarr

Kriegsfeld

2022

Biol Sex Differ

View full text Add to dashboard Cite

Despite recent work demonstrating that female rodents and humans do not show greater variance in behavior and physiology than males due to ovulatory cycles, many researchers still default to using males in their investigations. Although government funding agencies now require inclusion of female subjects where applicable, the erroneous belief that the study of males reduces overall data variance continues to result in male subject bias. Recently, we reported the first direct experimental refutation of this belief by examining continuous body temperature and locomotor activity in male and female mice. These findings revealed that males exceeded female variance within and across individuals over time, showing greater variance within a day than females do across an entire estrous cycle. However, the possibility remains that male variance within a day is impacted by ultradian rhythms, analogous to the influence of infradian estrous cycles on female variance, and both sexes show predictable, structured variance across the day. If structures underlying variance can be predicted, then the variance can be statistically accounted for, reducing experimental error and increasing precision of measurements. Here we assess these continuous body temperature and activity data for the contributions of structured and unstructured variance to overall variance within and across individuals at ultradian, circadian, and infradian timescales. In no instance do females exceed male variance, and in most instances male variance exceeds female variance. Additionally, more female variance is accounted for by temporal structure. In conclusion, even when estrous cycles are not controlled for, females show less variability than males, and this advantage can be further capitalized upon by inclusion of known temporal patterns to control for previously unknown but structured sources of variance.

show abstract

Section: Discussionmentioning

confidence: 99%

Female mice exhibit less overall variance, with a higher proportion of structured variance, than males at multiple timescales of continuous body temperature and locomotive activity records

Smarr

Kriegsfeld

2022

Biol Sex Differ

View full text Add to dashboard Cite

show abstract

“…This is the main reason to prefer metrics (such as the Euclidean distance) over measures (such as the warped distance). The research community is aware that DTW is a measure of distance rather than a metric, because DTW does not address the inequality of the triangle [ 31 , 32 ]. However, in the limit, as the warping constraint w approaches zero, DTW tends to be a metric [ 20 ].…”

Section: Methodsmentioning

confidence: 99%

The Accuracy of the Microsoft Kinect V2 Sensor for Human Gait Analysis. A Different Approach for Comparison with the Ground Truth

Guffanti

Brunete

Hernando

et al. 2020

Sensors

View full text Add to dashboard Cite

Several studies have examined the accuracy of the Kinect V2 sensor during gait analysis. Usually the data retrieved by the Kinect V2 sensor are compared with the ground truth of certified systems using a Euclidean comparison. Due to the Kinect V2 sensor latency, the application of a uniform temporal alignment is not adequate to compare the signals. On that basis, the purpose of this study was to explore the abilities of the dynamic time warping (DTW) algorithm to compensate for sensor latency (3 samples or 90 ms) and develop a proper accuracy estimation. During the experimental stage, six iterations were performed using the a dual Kinect V2 system. The walking tests were developed at a self-selected speed. The sensor accuracy for Euclidean matching was consistent with that reported in previous studies. After latency compensation, the sensor accuracy demonstrated considerably lower error rates for all joints. This demonstrated that the accuracy was underestimated due to the use of inappropriate comparison techniques. On the contrary, DTW is a potential method that compensates for the sensor latency, and works sufficiently in comparison with certified systems.

show abstract

“…Dynamic Time Warping (DTW) has been shown to be the best similarity measure to use along with the k-Nearest Neighbors (k-NN) [18]. DTW is not a distance metric as it does not fully satisfy the required properties (the triangle inequality in particular), but its use as similarity measure along with the NN-rule is valid [19]. There are two versions of kNN-DTW for MTS, dependent (DTW D ) and independent (DTW I ), and neither dominates the other [20].…”

Section: Mts Classifiersmentioning

confidence: 99%

XCM: An Explainable Convolutional Neural Network for Multivariate Time Series Classification

et al. 2021

View full text Add to dashboard Cite

Multivariate Time Series (MTS) classification has gained importance over the past decade with the increase in the number of temporal datasets in multiple domains. The current state-of-the-art MTS classifier is a heavyweight deep learning approach, which outperforms the second-best MTS classifier only on large datasets. Moreover, this deep learning approach cannot provide faithful explanations as it relies on post hoc model-agnostic explainability methods, which could prevent its use in numerous applications. In this paper, we present XCM, an eXplainable Convolutional neural network for MTS classification. XCM is a new compact convolutional neural network which extracts information relative to the observed variables and time directly from the input data. Thus, XCM architecture enables a good generalization ability on both large and small datasets, while allowing the full exploitation of a faithful post hoc model-specific explainability method (Gradient-weighted Class Activation Mapping) by precisely identifying the observed variables and timestamps of the input data that are important for predictions. We first show that XCM outperforms the state-of-the-art MTS classifiers on both the large and small public UEA datasets. Then, we illustrate how XCM reconciles performance and explainability on a synthetic dataset and show that XCM enables a more precise identification of the regions of the input data that are important for predictions compared to the current deep learning MTS classifier also providing faithful explainability. Finally, we present how XCM can outperform the current most accurate state-of-the-art algorithm on a real-world application while enhancing explainability by providing faithful and more informative explanations.

show abstract

Is the DTW “distance” really a metric? An algorithm reducing the number of DTW comparisons in isolated word recognition

Cited by 34 publications

References 7 publications

Female mice exhibit less overall variance, with a higher proportion of structured variance, than males at multiple timescales of continuous body temperature and locomotive activity records

Female mice exhibit less overall variance, with a higher proportion of structured variance, than males at multiple timescales of continuous body temperature and locomotive activity records

The Accuracy of the Microsoft Kinect V2 Sensor for Human Gait Analysis. A Different Approach for Comparison with the Ground Truth

XCM: An Explainable Convolutional Neural Network for Multivariate Time Series Classification

Contact Info

Product

Resources

About