Mining condensed sets of frequent episodes with more accurate frequencies from complex sequences

Many previous approaches to frequent episode discovery only accept simple sequences. Although a recent approach has been able to nd frequent episodes from complex sequences, the discovered sets are neither condensed nor accurate. This paper investigates the discovery of condensed sets of frequent episodes from complex sequences. We adopt a novel anti-monotonic frequency measure based on non-redundant occurrences, and dene a condensed set, nDaCF (the set of non-derivable approximately closed frequent episodes) within a given maximal error bound of support. We then introduce a series of effective pruning strategies, and develop a method, nDaCF-Miner, for discovering nDaCF sets. Experimental results show that, when the error bound is somewhat high, the discovered nDaCF sets are two orders of magnitude smaller than complete sets, and nDaCF-miner is more efficient than previous mining approaches. In addition, the nDaCF sets are more accurate than the sets found by previous approaches.

Recognition from appearance subspaces across image sets of variable scale

Linear subspace representations of appearance variation are pervasive in computer vision. In this paper we address the problem of robustly matching them (computing the similarity between them) when they correspond to sets of images of different (possibly greatly so) scales. We show that the naïve solution of projecting the low-scale subspace into the high-scale image space is inadequate, especially at large scale discrepancies. A successful approach is proposed instead. It consists of (i) an interpolated projection of the low-scale subspace into the high-scale space, which is followed by (ii) a rotation of this initial estimate within the bounds of the imposed “downsampling constraint”. The optimal rotation is found in the closed-form which best aligns the high-scale reconstruction of the low-scale subspace with the reference it is compared to. The proposed method is evaluated on the problem of matching sets of face appearances under varying illumination. In comparison to the naïve matching, our algorithm is shown to greatly increase the separation of between-class and within-class similarities, as well as produce far more meaningful modes of common appearance on which the match score is based.

A penalty-based aggregation operator for non-convex intervals

In the case of real-valued inputs, averaging aggregation functions have been studied extensively with results arising in fields including probability and statistics, fuzzy decision-making, and various sciences. Although much of the behavior of aggregation functions when combining standard fuzzy membership values is well established, extensions to interval-valued fuzzy sets, hesitant fuzzy sets, and other new domains pose a number of difficulties. The aggregation of non-convex or discontinuous intervals is usually approached in line with the extension principle, i.e. by aggregating all real-valued input vectors lying within the interval boundaries and taking the union as the final output. Although this is consistent with the aggregation of convex interval inputs, in the non-convex case such operators are not idempotent and may result in outputs which do not faithfully summarize or represent the set of inputs. After giving an overview of the treatment of non-convex intervals and their associated interpretations, we propose a novel extension of the arithmetic mean based on penalty functions that provides a representative output and satisfies idempotency.

On backwards and forwards reachable sets bounding for perturbed time-delay systems

Linear systems with interval time-varying delay and unknown-but-bounded disturbances are considered in this paper. We study the problem of finding outer bound of forwards reachable sets and inter bound of backwards reachable sets of the system. Firstly, two definitions on forwards and backwards reachable sets, where initial state vectors are not necessary to be equal to zero, are introduced. Then, by using the Lyapunov-Krasovskii method, two sufficient conditions for the existence of: (i) the smallest possible outer bound of forwards reachable sets; and (ii) the largest possible inter bound of backwards reachable sets, are derived. These conditions are presented in terms of linear matrix inequalities with two parameters need to tuned, which therefore can be efficiently solved by combining existing convex optimization algorithms with a two-dimensional search method to obtain optimal bounds. Lastly, the obtained results are illustrated by four numerical examples.

Reachable sets bounding for switched systems with time-varying delay and bounded disturbances

In this paper, we address the problem of finding outer bound of forward reachable sets and inter-bound of backward reachable sets of switched systems with an interval time-varying delay and bounded disturbances. By constructing a flexible Lyapunov–Krasovskii functional combining with some recent refined integral-based inequalities, some sufficient conditions are derived for the existence of (1) the smallest possible outer bound of forwards reachable sets; and (2) the largest possible inter-bound of backward reachable sets. These conditions are delay dependent and in the form of linear matrix inequalities, which therefore can be efficiently solved by using existing convex algorithms. A constructive geometric design of switching laws is also presented. Two numerical examples with simulation results are provided to demonstrate the effectiveness of our results.

Reachable sets bounding for generalized neural networks with interval time-varying delay and bounded disturbances

This paper deals with the problem of finding outer bound of forwards reachable sets and interbound of backwards reachable sets of generalized neural network systems with interval nondifferentiable time-varying delay and bounded disturbances. Based on constructing a suitable Lyapunov–Krasovskii functional and utilizing some improved Jensen integral-based inequalities, two sufficient conditions are derived for the existence of: (1) the smallest possible outer bound of forwards reachable sets and (2) the largest possible interbound of backwards reachable sets. These conditions are delay dependent and in the form of matrix inequalities, which therefore can be efficiently solved by using existing convex algorithms. Three numerical examples with simulation results are provided to demonstrate the effectiveness of our results.

Density based fuzzy c-means clustering of non-convex patterns

We propose a new technique to perform unsupervised data classification (clustering) based on density induced metric and non-smooth optimization. Our goal is to automatically recognize multidimensional clusters of non-convex shape. We present a modification of the fuzzy c-means algorithm, which uses the data induced metric, defined with the help of Delaunay triangulation. We detail computation of the distances in such a metric using graph algorithms. To find optimal positions of cluster prototypes we employ the discrete gradient method of non-smooth optimization. The new clustering method is capable to identify non-convex overlapped d-dimensional clusters.

Spatial and temporal changes in abundance of the infaunal bivalve Soletellina alba (Lamark, 1818) during a time of drought in the seasonally-closed Hopkins River Estuary, Victoria Australia

The infaunal bivalve Soletellina alba is susceptible to mass mortalities during annual winter flooding in the Hopkins River Estuary, southern Australia. Periods of low salinity (≤1) are the likely cause of these mass mortality events, which can occur in seasonally-closed estuaries when high winter flows are sufficient to flush all salt water from the estuary. Core samples of S. alba were collected from two water depths across four times and at three sites near the mouth of the estuary. Minimal to zero abundances of large S. alba (>1 mm) were expected to be sampled, particularly at the shallower water depth, during a typical winter flood event. However, the present study occurred during a period of drought, which led to the absence of winter flooding. This absence of winter flooding prevented the occurrence of lethal salinities (i.e. ≤1) in the estuary during this period and a greater number of living S. alba adults were sampled. Abundances of juvenile and adult S. alba were still variable, even in the absence of winter flooding, and reflected an interaction between date, site and water depth. However, no mass mortalities of adults were observed during the drought conditions in contrast to what occurs during typical winter flood events and provides support for the hypothesis that winter flooding is responsible for past mass mortalities.

Fuzzy clustering of non-convex patterns using global optimization

This paper discusses various extensions of the classical within-group sum of squared errors functional, routinely used as the clustering criterion. Fuzzy c-means algorithm is extended to the case when clusters have irregular shapes, by representing the clusters with more than one prototype. The resulting minimization problem is non-convex and non-smooth. A recently developed cutting angle method of global optimization is applied to this difficult problem

The need for low bias algorithms in classification learning from large data sets

This paper reviews the appropriateness for application to large data sets of standard machine learning algorithms, which were mainly developed in the context of small data sets. Sampling and parallelisation have proved useful means for reducing computation time when learning from large data sets. However, such methods assume that algorithms that were designed for use with what are now considered small data sets are also fundamentally suitable for large data sets. It is plausible that optimal learning from large data sets requires a different type of algorithm to optimal learning from small data sets. This paper investigates one respect in which data set size may affect the requirements of a learning algorithm — the bias plus variance decomposition of classification error. Experiments show that learning from large data sets may be more effective when using an algorithm that places greater emphasis on bias management, rather than variance management.

An optimal strategy for extracting probabilistic rules by combining rough sets and genetic algorithm

This paper proposes an optimal strategy for extracting probabilistic rules from databases. Two inductive learning-based statistic measures and their rough set-based definitions: accuracy and coverage are introduced. The simplicity of a rule emphasized in this paper has previously been ignored in the discovery of probabilistic rules. To avoid the high computational complexity of rough-set approach, some rough-set terminologies rather than the approach itself are applied to represent the probabilistic rules. The genetic algorithm is exploited to find the optimal probabilistic rules that have the highest accuracy and coverage, and shortest length. Some heuristic genetic operators are also utilized in order to make the global searching and evolution of rules more efficiently. Experimental results have revealed that it run more efficiently and generate probabilistic classification rules of the same integrity when compared with traditional classification methods.

The rough sets feature selection for trees recognition in color aerial images using genetic algorithms

Selecting a set of features which is optimal for a given task is the problem which plays an important role in a wide variety of contexts including pattern recognition, images understanding and machine learning. The concept of reduction of the decision table based on the rough set is very useful for feature selection. In this paper, a genetic algorithm based approach is presented to search the relative reduct decision table of the rough set. This approach has the ability to accommodate multiple criteria such as accuracy and cost of classification into the feature selection process and finds the effective feature subset for texture classification . On the basis of the effective feature subset selected, this paper presents a method to extract the objects which are higher than their surroundings, such as trees or forest, in the color aerial images. The experiments results show that the feature subset selected and the method of the object extraction presented in this paper are practical and effective.

The reduction subset based on rough sets applied to texture classification

The rough set is a new mathematical approach to imprecision, vagueness and uncertainty. The concept of reduction of the decision table based on the rough sets is very useful for feature selection. The paper describes an application of rough sets method to feature selection and reduction in texture images recognition. The methods applied include continuous data discretization based on Fuzzy c-means and, and rough set method for feature selection and reduction. The trees extractions in the aerial images were applied. The experiments show that the methods presented in this paper are practical and effective.

Convex hulls as an hypothesis language bias

Classification learning is dominated by systems which induce large numbers of small axis-orthogonal decision surfaces which biases such systems towards particular hypothesis types. However, there is reason to believe that many domains have underlying concepts which do not involve axis orthogonal surfaces. Further, the multiplicity of small decision regions mitigates against any holistic appreciation of the theories produced by these systems, notwithstanding the fact that many of the small regions are individually comprehensible. We propose the use of less strongly biased hypothesis languages which might be expected to model' concepts using a number of structures close to the number of actual structures in the domain. An instantiation of such a language, a convex hull based classifier, CHI, has been implemented to investigate modeling concepts as a small number of large geometric structures in n-dimensional space. A comparison of the number of regions induced is made against other well-known systems on a representative selection of largely or wholly continuous valued machine learning tasks. The convex hull system is shown to produce a number of induced regions about an order of magnitude less than well-known systems and very close to the number of actual concepts. This representation, as convex hulls, allows the possibility of extraction of higher level mathematical descriptions of the induced concepts, using the techniques of computational geometry.