Russel Pears

Followers

Following

Co-authors

Public Views

Interests

Uploads

Papers by Russel Pears

A Novel Evolving Clustering Algorithm with Polynomial Regression for Chaotic Time-Series Prediction

by Lukas Lukas and Russel Pears

Time-series prediction has been a very well researched topic in recent studies. Some popular appr... more Time-series prediction has been a very well researched topic in
recent studies. Some popular approaches to this problem are the traditional statistical methods e.g. multiple linear regression and moving average, and neural network with the Multi Layer Perceptron which has shown its supremacy in time-series prediction. In this study, we used a different approach based on
evolving clustering algorithm with polynomial regressions to find repeating local patterns in a time-series data. To illustrate chaotic time-series data we have taken into account the use of stock price data from Indonesian stock exchange market and currency exchange rate data. In addition, we have also
conducted a benchmark test using the Mackey Glass data set. Results showed that the algorithm offers a considerably high accuracy in time-series prediction and could also reveal repeating patterns of movement from the past.

Download

Measuring cascade effects in interdependent networks by using effective graph resistance

2015 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), 2015

Use of Ensembles of Fourier Spectra in Capturing Recurrent Concepts in Data Streams

In this research, we apply ensembles of Fourier encoded spectra to capture and mine recurring con... more In this research, we apply ensembles of Fourier encoded spectra to capture and mine recurring concepts in a data stream environment. Previous research showed that compact versions of Decision Trees can be obtained by applying the Discrete Fourier Transform to accurately capture recurrent concepts in a data stream. However, in highly volatile environments where new concepts emerge often, the approach of encoding each concept in a separate spectrum is no longer viable due to memory overload and thus in this research we present an ensemble approach that addresses this problem. Our empirical results on real world data and synthetic data exhibiting varying degrees of recurrence reveal that the ensemble approach outperforms the single spectrum approach in terms of classification accuracy, memory and execution time.

Download

Evaluating the Cascade Effect in Interdependent Networks via Algebraic Connectivity

Australasian Journal of Information, Communication Technology and Applications, 2015

ABSTRACT Understanding how the underlying network structure and interconnectivity impact on the r... more ABSTRACT Understanding how the underlying network structure and interconnectivity impact on the robustness of the interdependent networks is a major challenge in complex networks studies. There are some existing metrics that can be used to measure network robustness. However, different metrics such as the average node degree interprets different characteristic of network topological structure, especially less metrics have been identified to effectively evaluate the cascade performance in interdependent networks. In this paper, we propose to use a combined Laplacian matrix to model the interdependent networks and their interconnectivity, and then use its algebraic connectivity metric as a measure to evaluate its cascading behavior. Moreover, we have conducted extensive comparative studies among different metrics such as the average node degree, and the proposed algebraic connectivity. We have found that the algebraic connectivity metric can describe more accurate and finer characteristics on topological structure of the interdependent networks than other metrics widely adapted by the existing research studies for evaluating the cascading performance in interdependent networks.

Mining Recurrent Concepts in Data Streams Using the Discrete Fourier Transform

Lecture Notes in Computer Science, 2014

In this research we address the problem of capturing recurring concepts in a data stream environm... more In this research we address the problem of capturing recurring concepts in a data stream environment. Recurrence capture enables the re-use of previously learned classifiers without the need for re-learning while providing for better accuracy during the concept recurrence interval. We capture concepts by applying the Discrete Fourier Transform (DFT) to Decision Tree classifiers to obtain highly compressed versions of the trees at concept drift points in the stream and store such trees in a repository for future use. Our empirical results on real world and synthetic data exhibiting varying degrees of recurrence show that the Fourier compressed trees are more robust to noise and are able to capture recurring concepts with higher precision than a meta learning approach that chooses to re-use classifiers in their originally occurring form. The DFT, apart from its use in meta learning, has a number of other desirable properties that make it attractive for mining high speed data streams. This includes the ability to classify directly from the spectra generated, thus eliminating the need for expensive traversal of a tree structure.

Download

Searching Genomic Databases using the Prime Factor Filter

2006 International Conference on Information and Automation, 2006

The major bottleneck in searching genomic databases is the sheer size of the databases involved. ... more The major bottleneck in searching genomic databases is the sheer size of the databases involved. A number of different solutions to the problem of aligning query sequences to genomic databases have been proposed, including the widely used BLAST and FASTA systems. While such systems are effective against traditional applications such as query alignment, they do not scale well for applications such as whole genome shotgun sequencing and all versus all comparisons of one organism against another. The latter application has quadratic time complexity in the size of the databases involved and requires a different approach to BLAST type search engines that rely on a linear scan of the database. Our approach relies on a two-stage filter to prune a significant fraction of the database prior to alignment. The filter uses the MRS index[8] as the first stage followed by a novel indexing scheme that we propose in this paper. The MRS index screens sequences that map to the same frequency vector and has been shown to produce speedups of up to 12 over systems that do not employ such an index. However, the MRS index is inadequate against sequences that are inherently different while still mapping to the same frequency vector. Our filter, based on the prime factor Indexing scheme is successful in eliminating a large fraction of such false positives that survive the MRS index. Our experiments show that at least 75% of the false positives is eliminated, resulting in speedups of up to 5 times over the MRS indexing scheme.

Boosting Prediction Accuracy of Bad Payments in Financial Credit Applications

Technologies for Infrequent and Critical Event Detection, 2010

Personalised modelling for multiple time-series data prediction

The behaviour of multiple stock markets can be described within the framework of complex dynamic ... more The behaviour of multiple stock markets can be described within the framework of complex dynamic systems (CDS). Using a global model with the Kalman Filter we are able to extract the dynamic interaction network (DIN) of these markets. The model was shown to successfully capture interactions between stock markets in the long term. In this study we investigate the effectiveness

Accelerating multi dimensional queries in data warehouses

Processing or OLAP is the main vehicle for querying data warehouses. OLAP operations commonly inv... more Processing or OLAP is the main vehicle for querying data warehouses. OLAP operations commonly involve the computation of multidimensional aggregates. The major bottleneck in computing these aggregates is the large volume of data that needs to be processed which in turn leads to prohibitively expensive query execution times. On the other hand, Data Analysts are primarily concerned with discerning trends in the data and thus a system that provides approximate answers in a timely fashion would suit their requirements better.

Download

A Multi-Methodological Approach to Rare Association Rule Mining

Non-redundant rare itemset generation

How effective is model checking in practice?

Collaborative Wideband Compressed Signal Detection in Interplanetary Internet

by Russel Pears and S. Sakthithasan

Dynamic Symbolic Execution Guided by Data Dependency Analysis for High Structural Coverage

Tracking Drift Types in Changing Data Streams

Ontologies and machine learning systems

by Russel Pears and S. Tegginmath

Web Materialization Formulation: Modelling Feasible Solutions

Mining developer communication streams