arXiv subjects

Anthony Bagnall

Publications and source records attributed to Anthony Bagnall.

34 records · Page 2Linked to original sources

arXiv2019

A tale of two toolkits, report the first: benchmarking time series classification algorithms for correctness and efficiency

sktime is an open source, Python based, sklearn compatible toolkit for time series analysis developed by researchers at the University of East Anglia (UEA), University College London and the Alan Turing Institute. A key initial goal for sktime was to provide time series classification functionality equivalent to that available in a related java package, tsml, also developed at UEA. We describe the implementation of six such classifiers in sktime and compare them to their tsml equivalents. We demonstrate correctness through equivalence of accuracy on a range of standard test problems and compare the build time of the different implementations. We find that there is significant difference in accuracy on only one of the six algorithms we look at (Proximity Forest). This difference is causing us some pain in debugging. We found a much wider range of difference in efficiency. Again, this was not unexpected, but it does highlight ways both toolkits could be improved.

cs.LG↗

arXiv2019

sktime: A Unified Interface for Machine Learning with Time Series

We present sktime -- a new scikit-learn compatible Python library with a unified interface for machine learning with time series. Time series data gives rise to various distinct but closely related learning tasks, such as forecasting and time series classification, many of which can be solved by reducing them to related simpler tasks. We discuss the main rationale for creating a unified interface, including reduction, as well as the design of sktime's core API, supported by a clear overview of common time series tasks and reduction approaches.

cs.LG↗

arXiv2019

The UCR Time Series Archive

The UCR Time Series Archive - introduced in 2002, has become an important resource in the time series data mining community, with at least one thousand published papers making use of at least one data set from the archive. The original incarnation of the archive had sixteen data sets but since that time, it has gone through periodic expansions. The last expansion took place in the summer of 2015 when the archive grew from 45 to 85 data sets. This paper introduces and will focus on the new data expansion from 85 to 128 data sets. Beyond expanding this valuable resource, this paper offers pragmatic advice to anyone who may wish to evaluate a new algorithm on the archive. Finally, this paper makes a novel and yet actionable claim: of the hundreds of papers that show an improvement over the standard baseline (1-nearest neighbor classification), a large fraction may be mis-attributing the reasons for their improvement. Moreover, they may have been able to achieve the same improvement with a much simpler modification, requiring just a single line of code.

cs.LG↗

arXiv2019

Scalable Dictionary Classifiers for Time Series Classification

Dictionary based classifiers are a family of algorithms for time series classification (TSC), that focus on capturing the frequency of pattern occurrences in a time series. The ensemble based Bag of Symbolic Fourier Approximation Symbols (BOSS) was found to be a top performing TSC algorithm in a recent evaluation, as well as the best performing dictionary based classifier. A recent addition to the category, the Word Extraction for Time Series Classification (WEASEL), claims an improvement on this performance. Both of these algorithms however have non-trivial scalability issues, taking a considerable amount of build time and space on larger datasets. We evaluate changes to the way BOSS chooses classifiers for its ensemble, replacing its parameter search with random selection. This change allows for the easy implementation of contracting, setting a build time limit for the classifier and check-pointing, saving progress during the classifiers build. To differentiate between the two BOSS ensemble methods we refer to our randomised version as RBOSS. Additionally we test the application of common ensembling techniques to help retain accuracy from the loss of the BOSS parameter search. We achieve a significant reduction in build time without a significant change in accuracy on average when compared to BOSS by creating a size $n$ weighted ensemble selecting the best performers from $k$ randomly chosen parameter sets. Our experiments are conducted on datasets from the recently expanded UCR time series archive. We demonstrate the usability improvements to RBOSS with a case study using a large whale acoustics dataset for which BOSS proved infeasible.

cs.LG↗

arXiv2018

Can automated smoothing significantly improve benchmark time series classification algorithms?

tl;dr: no, it cannot, at least not on average on the standard archive problems. We assess whether using six smoothing algorithms (moving average, exponential smoothing, Gaussian filter, Savitzky-Golay filter, Fourier approximation and a recursive median sieve) could be automatically applied to time series classification problems as a preprocessing step to improve the performance of three benchmark classifiers (1-Nearest Neighbour with Euclidean and Dynamic Time Warping distances, and Rotation Forest). We found no significant improvement over unsmoothed data even when we set the smoothing parameter through cross validation. We are not claiming smoothing has no worth. It has an important role in exploratory analysis and helps with specific classification problems where domain knowledge can be exploited. What we observe is that the automatic application does not help and that we cannot explain the improvement of other time series classification algorithms over the baseline classifiers simply as a function of the absence of smoothing.

cs.LG↗

arXiv2018

The UEA multivariate time series classification archive, 2018

In 2002, the UCR time series classification archive was first released with sixteen datasets. It gradually expanded, until 2015 when it increased in size from 45 datasets to 85 datasets. In October 2018 more datasets were added, bringing the total to 128. The new archive contains a wide range of problems, including variable length series, but it still only contains univariate time series classification problems. One of the motivations for introducing the archive was to encourage researchers to perform a more rigorous evaluation of newly proposed time series classification (TSC) algorithms. It has worked: most recent research into TSC uses all 85 datasets to evaluate algorithmic advances. Research into multivariate time series classification, where more than one series are associated with each class label, is in a position where univariate TSC research was a decade ago. Algorithms are evaluated using very few datasets and claims of improvement are not based on statistical comparisons. We aim to address this problem by forming the first iteration of the MTSC archive, to be hosted at the website www.timeseriesclassification.com. Like the univariate archive, this formulation was a collaborative effort between researchers at the University of East Anglia (UEA) and the University of California, Riverside (UCR). The 2018 vintage consists of 30 datasets with a wide range of cases, dimensions and series lengths. For this first iteration of the archive we format all data to be of equal length, include no series with missing data and provide train/test splits.

cs.LG↗

arXiv2018

From BOP to BOSS and Beyond: Time Series Classification with Dictionary Based Classifiers

A family of algorithms for time series classification (TSC) involve running a sliding window across each series, discretising the window to form a word, forming a histogram of word counts over the dictionary, then constructing a classifier on the histograms. A recent evaluation of two of this type of algorithm, Bag of Patterns (BOP) and Bag of Symbolic Fourier Approximation Symbols (BOSS) found a significant difference in accuracy between these seemingly similar algorithms. We investigate this phenomenon by deconstructing the classifiers and measuring the relative importance of the four key components between BOP and BOSS. We find that whilst ensembling is a key component for both algorithms, the effect of the other components is mixed and more complex. We conclude that BOSS represents the state of the art for dictionary based TSC. Both BOP and BOSS can be classed as bag of words approaches. These are particularly popular in Computer Vision for tasks such as image classification. Converting approaches from vision requires careful engineering. We adapt three techniques used in Computer Vision for TSC: Scale Invariant Feature Transform; Spatial Pyramids; and Histrogram Intersection. We find that using Spatial Pyramids in conjunction with BOSS (SP) produces a significantly more accurate classifier. SP is significantly more accurate than standard benchmarks and the original BOSS algorithm. It is not significantly worse than the best shapelet based approach, and is only outperformed by HIVE-COTE, an ensemble that includes BOSS as a constituent module.

cs.LG↗

arXiv2017

A Shapelet Transform for Multivariate Time Series Classification

Shapelets are phase independent subsequences designed for time series classification. We propose three adaptations to the Shapelet Transform (ST) to capture multivariate features in multivariate time series classification. We create a unified set of data to benchmark our work on, and compare with three other algorithms. We demonstrate that multivariate shapelets are not significantly worse than other state-of-the-art algorithms.

cs.LG↗

arXiv2017

The Heterogeneous Ensembles of Standard Classification Algorithms (HESCA): the Whole is Greater than the Sum of its Parts

Building classification models is an intrinsically practical exercise that requires many design decisions prior to deployment. We aim to provide some guidance in this decision making process. Specifically, given a classification problem with real valued attributes, we consider which classifier or family of classifiers should one use. Strong contenders are tree based homogeneous ensembles, support vector machines or deep neural networks. All three families of model could claim to be state-of-the-art, and yet it is not clear when one is preferable to the others. Our extensive experiments with over 200 data sets from two distinct archives demonstrate that, rather than choose a single family and expend computing resources on optimising that model, it is significantly better to build simpler versions of classifiers from each family and ensemble. We show that the Heterogeneous Ensembles of Standard Classification Algorithms (HESCA), which ensembles based on error estimates formed on the train data, is significantly better (in terms of error, balanced error, negative log likelihood and area under the ROC curve) than its individual components, picking the component that is best on train data, and a support vector machine tuned over 1089 different parameter configurations. We demonstrate HESCA+, which contains a deep neural network, a support vector machine and two decision tree forests, is significantly better than its components, picking the best component, and HESCA. We analyse the results further and find that HESCA and HESCA+ are of particular value when the train set size is relatively small and the problem has multiple classes. HESCA is a fast approach that is, on average, as good as state-of-the-art classifiers, whereas HESCA+ is significantly better than average and represents a strong benchmark for future research.

cs.LG↗

arXiv2017

Simulated Data Experiments for Time Series Classification Part 1: Accuracy Comparison with Default Settings

There are now a broad range of time series classification (TSC) algorithms designed to exploit different representations of the data. These have been evaluated on a range of problems hosted at the UCR-UEA TSC Archive (www.timeseriesclassification.com), and there have been extensive comparative studies. However, our understanding of why one algorithm outperforms another is still anecdotal at best. This series of experiments is meant to help provide insights into what sort of discriminatory features in the data lead one set of algorithms that exploit a particular representation to be better than other algorithms. We categorise five different feature spaces exploited by TSC algorithms then design data simulators to generate randomised data from each representation. We describe what results we expected from each class of algorithm and data representation, then observe whether these prior beliefs are supported by the experimental evidence. We provide an open source implementation of all the simulators to allow for the controlled testing of hypotheses relating to classifier performance on different data representations. We identify many surprising results that confounded our expectations, and use these results to highlight how an over simplified view of classifier structure can often lead to erroneous prior beliefs. We believe ensembling can often overcome prior bias, and our results support the belief by showing that the ensemble approach adopted by the Hierarchical Collective of Transform based Ensembles (HIVE-COTE) is significantly better than the alternatives when the data representation is unknown, and is significantly better than, or not significantly significantly better than, or not significantly worse than, the best other approach on three out of five of the individual simulators.

cs.LG↗

arXiv2017

On the Use of Default Parameter Settings in the Empirical Evaluation of Classification Algorithms

We demonstrate that, for a range of state-of-the-art machine learning algorithms, the differences in generalisation performance obtained using default parameter settings and using parameters tuned via cross-validation can be similar in magnitude to the differences in performance observed between state-of-the-art and uncompetitive learning systems. This means that fair and rigorous evaluation of new learning algorithms requires performance comparison against benchmark methods with best-practice model selection procedures, rather than using default parameter settings. We investigate the sensitivity of three key machine learning algorithms (support vector machine, random forest and rotation forest) to their default parameter settings, and provide guidance on determining sensible default parameter values for implementations of these algorithms. We also conduct an experimental comparison of these three algorithms on 121 classification problems and find that, perhaps surprisingly, rotation forest is significantly more accurate on average than both random forest and a support vector machine.

cs.LG↗

arXiv2016

The Great Time Series Classification Bake Off: An Experimental Evaluation of Recently Proposed Algorithms. Extended Version

In the last five years there have been a large number of new time series classification algorithms proposed in the literature. These algorithms have been evaluated on subsets of the 47 data sets in the University of California, Riverside time series classification archive. The archive has recently been expanded to 85 data sets, over half of which have been donated by researchers at the University of East Anglia. Aspects of previous evaluations have made comparisons between algorithms difficult. For example, several different programming languages have been used, experiments involved a single train/test split and some used normalised data whilst others did not. The relaunch of the archive provides a timely opportunity to thoroughly evaluate algorithms on a larger number of datasets. We have implemented 18 recently proposed algorithms in a common Java framework and compared them against two standard benchmark classifiers (and each other) by performing 100 resampling experiments on each of the 85 datasets. We use these results to test several hypotheses relating to whether the algorithms are significantly more accurate than the benchmarks and each other. Our results indicate that only 9 of these algorithms are significantly more accurate than both benchmarks and that one classifier, the Collective of Transformation Ensembles, is significantly more accurate than all of the others. All of our experiments and results are reproducible: we release all of our code, results and experimental details and we hope these experiments form the basis for more rigorous testing of new algorithms in the future.

cs.LG↗

arXiv2014

Ensembles of Random Sphere Cover Classifiers

We propose and evaluate alternative ensemble schemes for a new instance based learning classifier, the Randomised Sphere Cover (RSC) classifier. RSC fuses instances into spheres, then bases classification on distance to spheres rather than distance to instances. The randomised nature of RSC makes it ideal for use in ensembles. We propose two ensemble methods tailored to the RSC classifier; $αβ$RSE, an ensemble based on instance resampling and $α$RSSE, a subspace ensemble. We compare $αβ$RSE and $α$RSSE to tree based ensembles on a set of UCI datasets and demonstrates that RSC ensembles perform significantly better than some of these ensembles, and not significantly worse than the others. We demonstrate via a case study on six gene expression data sets that $α$RSSE can outperform other subspace ensemble methods on high dimensional data when used in conjunction with an attribute filter. Finally, we perform a set of Bias/Variance decomposition experiments to analyse the source of improvement in comparison to a base classifier.

cs.LG↗

arXiv2014

Finding Motif Sets in Time Series

Time-series motifs are representative subsequences that occur frequently in a time series; a motif set is the set of subsequences deemed to be instances of a given motif. We focus on finding motif sets. Our motivation is to detect motif sets in household electricity-usage profiles, representing repeated patterns of household usage. We propose three algorithms for finding motif sets. Two are greedy algorithms based on pairwise comparison, and the third uses a heuristic measure of set quality to find the motif set directly. We compare these algorithms on simulated datasets and on electricity-usage data. We show that Scan MK, the simplest way of using the best-matching pair to find motif sets, is less accurate on our synthetic data than Set Finder and Cluster MK, although the latter is very sensitive to parameter settings. We qualitatively analyse the outputs for the electricity-usage data and demonstrate that both Scan MK and Set Finder can discover useful motif sets in such data.

cs.LG↗

arXiv2014

An Experimental Evaluation of Nearest Neighbour Time Series Classification

Data mining research into time series classification (TSC) has focussed on alternative distance measures for nearest neighbour classifiers. It is standard practice to use 1-NN with Euclidean or dynamic time warping (DTW) distance as a straw man for comparison. As part of a wider investigation into elastic distance measures for TSC~\cite{lines14elastic}, we perform a series of experiments to test whether this standard practice is valid. Specifically, we compare 1-NN classifiers with Euclidean and DTW distance to standard classifiers, examine whether the performance of 1-NN Euclidean approaches that of 1-NN DTW as the number of cases increases, assess whether there is any benefit of setting $k$ for $k$-NN through cross validation whether it is worth setting the warping path for DTW through cross validation and finally is it better to use a window or weighting for DTW. Based on experiments on 77 problems, we conclude that 1-NN with Euclidean distance is fairly easy to beat but 1-NN with DTW is not, if window size is set through cross validation.

cs.LG↗

arXiv2014

Predictive Modelling of Bone Age through Classification and Regression of Bone Shapes

Bone age assessment is a task performed daily in hospitals worldwide. This involves a clinician estimating the age of a patient from a radiograph of the non-dominant hand. Our approach to automated bone age assessment is to modularise the algorithm into the following three stages: segment and verify hand outline; segment and verify bones; use the bone outlines to construct models of age. In this paper we address the final question: given outlines of bones, can we learn how to predict the bone age of the patient? We examine two alternative approaches. Firstly, we attempt to train classifiers on individual bones to predict the bone stage categories commonly used in bone ageing. Secondly, we construct regression models to directly predict patient age. We demonstrate that models built on summary features of the bone outline perform better than those built using the one dimensional representation of the outline, and also do at least as well as other automated systems. We show that models constructed on just three bones are as accurate at predicting age as expert human assessors using the standard technique. We also demonstrate the utility of the model by quantifying the importance of ethnicity and sex on age development. Our conclusion is that the feature based system of separating the image processing from the age modelling is the best approach for automated bone ageing, since it offers flexibility and transparency and produces accurate estimates.

cs.LG↗