Many algorithms have been proposed for the problem of time series classification. However, it is clear that one-nearest-neighbor with Dynamic Time Warping (DTW) distance exceptionally difficult to beat. This approach has one weakness, however; computationally too demanding many realtime applications. One way mitigate this speed up DTW calculations. Nonetheless, there a limit how much can help. In work, we propose an additional technique, numerosity reduction, DTW. While idea reduction...

The matching of two-dimensional shapes is an important problem with applications in domains as diverse biometrics, industry, medicine and anthropology. distance measure used must be invariant to many distortions, including scale, offset, noise, partial occlusion, etc. Most these distortions are relatively easy handle, either the representation data or similarity used. However rotation invariance seems uniquely difficult. Current approaches typically try achieve data, at expense...

For many real world problems we must perform classification under widely varying amounts of computational resources. example, if asked to classify an instance taken from a bursty stream, may have milliseconds minutes return class prediction. such anytime algorithm be especially useful. In this work show how can convert the ubiquitous nearest neighbor classifier into that produce instant classification, or given luxury additional time, utilize extra time increase accuracy. We demonstrate...

Fault diagnosis is the key technology to guarantee reliability and safety of satellite attitude control systems. Although model-based methods have achieved good fault performance, various factors (such as closed-loop design, model nonlinearity, external disturbances) in systems still bring challenges isolation. Meanwhile, data-driven can assist for based on residual signals (generated from methods) input/output data relieve difficulty. Based above motivations, a novel dual-driven approach...

Over the past three decades, there has been a great deal of research on shape analysis, focusing mostly indexing, clustering, and classification. In this work, we introduce new problem finding discords, most unusual shapes in collection. We motivate by considering utility discords diverse domains including zoology, anthropology, medicine. While brute force search algorithm quadratic time complexity, avoid using locality-sensitive hashing to estimate similarity between which enables us...

A prerequisite for the existing remaining useful life prediction methods based on stochastic processes is assumption of independent increments. However, this in sharp contrast to some practical systems including batteries and blast furnace walls, which degradation have property long-range dependence. Based fractional Brownian motion, we adopt a process with dependence predict above systems. Because neither Markovian nor semimartingale, exact analytical first passage time difficult derive...

Some practical systems such as blast furnaces and turbofan engines have degradation processes with memory effects. The term of effects implies that the future states depend on both current state past because interaction environments. However, most works generally used a memoryless Markovian process to model processes. To characterize in systems, we develop new type model, which diffusion is represented fractional Brownian motion (FBM). FBM actually special non-Markovian long-term...

The problem of efficiently finding images that are similar to a target image has attracted much attention in the processing community and is rightly considered an information retrieval task. However, structure regularities large datasets area which data mining beginning make fundamental contributions. In this work, we consider new discovering shape motifs, approximately repeated shapes within (or between) collections. As shall show, motifs can have applications tasks as diverse anthropology,...

The vast majority of visualization tools introduced so far are specialized pieces software that run explicitly on a particular dataset at time for purpose. In this work we introduce novel framework allowing to take place in the background normal day-to-day operation any GUI based system. Our system works by replacing standard file icons with automatically created reflect contents files principled way. We call such Intelligent Icons. utility Icons is further enhanced arranging them way...

Abstract Multivariable stochastic degradation system (MSDS) is quite common in industries such as blast furnace ironmaking, vehicle transportation, and aerospace manufacturing. Large‐scale complex equipments may be affected by multiple factors, resulting not just a single deteriorating performance characteristic. It difficult to handle unknown failure structures of practical systems using traditional univariate modeling methods. A novel health index (HI) constructed quantitatively analyze...

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> Object recognition and content-based image retrieval systems rely heavily on the accurate efficient identification of 2-D shapes. Features such as color, texture, positioning etc., are insufficient to convey information that could be obtained through shape analysis. A fundamental requirement in this analysis is similarities computed invariantly basic geometric transformations, e.g., scaling,...

Abstract Degradation processes of practical chemical engineering systems are difficult to model accurately because complicated nonlinearities, non‐stationarities, and non‐Markovian properties. Traditional prognostics techniques tend neglect the multi‐mode switching issue, therefore cannot be used for predicting remaining useful life (RUL) a piece long‐life equipment, especially when it is continuously operated under varying modes. Specifically, stationarity differential data also plays an...

Object recognition and content-based image retrieval systems rely heavily on the accurate efficient identification of shapes. A fundamental requirement in shape analysis process is that similarities should be computed invariantly to basic geometric transformations, e.g. scaling, shifting, most importantly, rotations. And while scale shift invariance are easily achievable through a suitable representation, rotation much harder deal with.In this work we explore metric properties invariant...

The performance of random forest (RF) based satellite attitude control system (ACS) fault diagnosis methods is limited by uninformative features in high-dimensional data. To solve this problem, we proposed a feature-weighted with Boruta (FWRFB) method for ACSs. Firstly, feature selection algorithm used to obtain set and determine significant weights. Subsequently, novel (FWRF) designed, which utilizes sampling instead simple generate subsets the RF. FWRFB effectively information while...