One of the contemporary challenges in anomaly detection is the ability to detect, and differentiate between, both point and collective anomalies within a data sequence or time series. The anomaly package has been developed to provide users with a choice of anomaly detection methods and, in particular, provides an implementation of the recently proposed Collective And Point Anomaly family of anomaly detection algorithms. This article describes the methods implemented whilst also highlighting their application to simulated data as well as real data examples contained in the package.

Changepoint detection is an important problem with a wide range of applications. There are many different types of changes that one may wish to detect, and a wide range of algorithms and software for detecting them. However there are relatively few approaches for detecting changes-in-slope in the mean of a signal plus noise model. We describe the R package cpop, available on the Comprehensive R Archive Network (CRAN). This package implements CPOP, a dynamic programming algorithm, to find the optimal set of changes that minimizes an L0 penalized cost, with the cost being a weighted residual sum of squares. The package has extended the CPOP algorithm so it can analyse data that is unevenly spaced, allow for heterogeneous noise variance, and allows for a grid of potential change locations to be different from the locations of the data points. There is also an implementation that uses the CROPS algorithm to detect all segmentations that are optimal as you vary the L0 penalty for adding a change across a continuous range of values.

Implements a series of robust Kalman filtering approaches. It implements the additive outlier robust filters of Ruckdeschel et al. (2014) and Agamennoni et al. (2018) , the innovative outlier robust filter of Ruckdeschel et al. (2014) , as well as the innovative and additive outlier robust filter of Fisch et al. (2020) .

There exist a large number of computationally intensive statistical procedures that can be implemented in a manner that is suitable for evaluation using a parallel computing environment. Within this number there exists a class of procedures, often described as “course grained parallel” or “embarrassingly parallel”. The defining characteristic of these procedures is that they can be reduced to a number of sub-procedures that are independent of each other and require little or no inter-procedure communication i.e. they can be executed concurrently. Initially, it might be thought that this class is too small to warrant significant attention, however this is far from being the case. For example, methodologies such as bootstrapping, cross-validation, many types of Markov Processes (including MCMC), and certain optimisation and search algorithms are of this type. Importantly, the increase in availability of High Throughput Computing (HTC) environments, consisting of large numbers of interconnected computers, has made employing such procedures particularly attractive, leading to a significant increase in the amount of research being undertaken using HTC, notably in the areas of biochemistry, genetics, pharmaceuticals, economics, financial modelling and the social sciences. A High Throughput Computing environment provides a means for processing a large number of independent (non-interacting) tasks simultaneously. In the simplest case, the HTC environment may employ only a single multi-processor system. At the other extreme, the HTC environment might comprise a large number of systems with different operating systems and hardware located across a number of different institutions and administrative domains. When this is the case the environment may be said to provide High Throughput Distributed Computing (HTDC). HTC on a single multiprocessor system is relatively straightforward. Typically the user has an account on the system (can be identified to the system by a user name and password) and can submit the tasks for processing by using the software tools available on that system. Higher level means of submitting tasks exist, such as the snow package for R [1]. This package allows functions defined in R or installed R packages to be invoked multiple times with varying argument signatures and executed on a number of processors simultaneously. In [1] it is noted that the functionality offered by snow could be extended to use the GRID, which by its nature provides a HTDC environment. Some of these extensions have been addressed within the GridR system [3], which is similar in principle to snow but provides some of the technical requirements necessary

Formal verification has become an important step in circuit and system design.A prominent technique is Bounded Model Checking (BMC) which is widely used in industry.In BMC it is checked if certain properties hold for the design.But even if all properties could be successfully verified, it is difficult to determine if the properties cover the entire functional behavior of the circuit.Recently, a new approach for estimating coverage in BMC has been presented that can easily be integrated in existing BMC tools.In this paper we give experimental results on the application of the technique to the block-level verification of a RISC CPU.The experiments show that the costs for coverage estimation are comparable to the verification costs.Furthermore it is demonstrated how the technique can be applied to achieve full coverage on a higher level.As an example, we investigate the instruction set verification of a RISC CPU.

The application of high performance computing to spatial analysis has long been of interest to geographical scientists and has spearheaded research in computations. Of particular note is the pioneering work undertaken by Stan Openshaw at the University of Newcastle and at the Centre for Computational Geography at Leeds University, of which an exemplar is the Geographical Analysis Machine [11]. More recently, Martin [10] has identified the potential for geocomputation to develop under the rubric of high performance computer (grid) networks and e-(electronic) social science. He identifies four essential research issues for e-social science: automated data mining; visualization of spatial data uncertainty; incorporation of an explicitly spatial dimension into simulation modelling; and neighbourhood classification from multi-source distributed datasets. Missing, perhaps, from Martin’s list is the use of computational grid to ’speed up’ the repetitious processes of many spatial statistics. What GAM, GWR (see below) and other methods of spatially localized analysis have in common is a general sequence of