Our shapelet transform was evaluated on 29 data sets. 18 of these came from the UCR Repository and can be downloaded from there. The 11 data sets we have created ourselves can be downloaded from Shapelet Transform data (ZIP, 14 MB). All problems have a train and test split.  

The zip file is password protected. To get the password, please read the points below, and if you agree to them, email Tony (ajb@uea.ac.uk),

  • Do not share the password with others.
  • If you are a postgraduate student/post-doc, you must discuss this with your supervisor first and CC him/her when requesting the password.
  • If you use the data, please reference

Hills, J, Lines, J, Baranauskas, E, Mapp, J, and Bagnall, A, Time Series Classification with Shapelets, Journal of Data Mining and Knowledge Discovery. online first preprint.pdf, 2013.

Bone Outline Classification

These data sets were generated as part of a project on automated bone age assessment.

The classification problem is to determine age group based on the outline of the bone.  Each bone is one of the proximal, distal or middle phalange of the thumb, middle finger or little finger. The data is arranged as eight separate TSC problems, although it is possible to treat it as one single multivariate TSC problem.

Problems: DP = Distal phalange.

DP_Little, DP_Middle, DP_Thumb, MP_Little, MP_Middle, PP_Little, PP_Middle, PP_Thumb

Otolith Outline Classification

Otoliths are calcium carbonate structures present in many vertebrates, found
within the sacculus of the pars inferior. There are three types of otoliths:
sagittae, lapilli, and asterisci. In fish, it is primarily the sagittal otoliths that are studied, as they are larger and easier to prepare and observe. Otoliths vary markedly in shape and size between species, but are of similar shape to other stocks of the same species. Otoliths contain information that can be used by `expert readers' to determine several key factors important in managing fish stock. Analysis of otolith boundaries may allow estimation of stock composition, including whether the samples are from one stock or multiple stocks, allowing management decisions to be made. We consider the problem of classifying herring stock (either North sea or Thames) based on the otolith outline

Problem Files: Herring_TRAIN.arff, Herring_TEST.arff

MPEG7 Image Outline Classification

MPEG-7 CE Shape-1 Part B is a database of binary images developed for testing MPEG-7 shape descriptors, and is available free online. It is used for testing contour/image and skeleton-based descriptors. Classes of images vary broadly, and include classes that are similar in shape to one another. There are 20 instances of each class, and 60 classes in total. We have extracted the outlines of these images and mapped them into 1-D series. We have created two time-series classication problems from the shapes, Beetle/Fly and Bird/Chicken.

Problem Files: BeetleFly_TRAIN.arff, BeetleFly _TEST.arff, BirdChicken_TRAIN.arff, BirdChicken _TEST.arff