Data Mining Efforts Research Articles

An Approach to Data Reduction and Integrated Machine Classification

The goal of the paper is to propose a novel approach to integrated machine classification and to investigate the effect of integration of the data reduction with data mining stage. The integration of both important steps of knowledge discovery in databases is recognized as a vital step towards improving effectiveness of the data mining effort. After having the introduced data reduction and integration schemes a solution to the integrated classification problem is proposed. The proposed algorithm allows for integrating data reduction through simultaneous instance and feature selection, with learning process using population-based and A-Team techniques. To validate the proposed approach and to investigate the effect of data reduction combined with different integration schemes, the computation experiment has been carried out. Experiment based on several benchmark datasets has shown that integrated data reduction and classifier learning outperform traditional approaches.

NCBI GEO: archive for high-throughput functional genomic data

The Gene Expression Omnibus (GEO) at the National Center for Biotechnology Information (NCBI) is the largest public repository for high-throughput gene expression data. Additionally, GEO hosts other categories of high-throughput functional genomic data, including those that examine genome copy number variations, chromatin structure, methylation status and transcription factor binding. These data are generated by the research community using high-throughput technologies like microarrays and, more recently, next-generation sequencing. The database has a flexible infrastructure that can capture fully annotated raw and processed data, enabling compliance with major community-derived scientific reporting standards such as ‘Minimum Information About a Microarray Experiment’ (MIAME). In addition to serving as a centralized data storage hub, GEO offers many tools and features that allow users to effectively explore, analyze and download expression data from both gene-centric and experiment-centric perspectives. This article summarizes the GEO repository structure, content and operating procedures, as well as recently introduced data mining features. GEO is freely accessible at http://www.ncbi.nlm.nih.gov/geo/.

Pattern of the Divergence of Olfactory Receptor Genes during Tetrapod Evolution

The olfactory receptor (OR) multigene family is responsible for the sense of smell in vertebrate species. OR genes are scattered widely in our chromosomes and constitute one of the largest gene families in eutherian genomes. Some previous studies revealed that eutherian OR genes diverged mainly during early mammalian evolution. However, the exact period when, and the ecological reason why eutherian ORs strongly diverged has remained unclear. In this study, I performed a strict data mining effort for marsupial opossum OR sequences and bootstrap analyses to estimate the periods of chromosomal migrations and gene duplications of OR genes during tetrapod evolution. The results indicate that chromosomal migrations occurred mainly during early vertebrate evolution before the monotreme-placental split, and that gene duplications occurred mainly during early mammalian evolution between the bird-mammal split and marsupial-placental split, coinciding with the reduction of opsin genes in primitive mammals. It could be thought that the previous chromosomal dispersal allowed the OR genes to subsequently expand easily, and the nocturnal adaptation of early mammals might have triggered the OR gene expansion.

Integrin‐Mediated Adhesion and Signaling in the Neurovascular Niche

Neural progenitors (NPs) in the central nervous system (CNS) reside in a neurovascular niche. Cell‐cell and cell‐extracellular matrix interactions within the neurovascular niche regulate NP self‐renewal and differentiation; however, the mechanisms used by NPs to communicate with blood vessels remain largely uncharacterized. We have shown previously that genetic ablation of αvβ8 integrin expression in NPs leads to profound neurovascular phenotypes due to defective cell‐ECM adhesion in the neurovascular niche. Here, we have analyzed αvβ8 integrin in NPs purified from wild type or αv and β8 integrin knockout (KO) mice. Cells lacking αvβ8 integrin display reduced self‐renewal, as assayed by neurosphere formation. Furthermore, αvβ8 KO cells display preferential differentiation toward the astroglial lineage following serum exposure. Subventricular zone NPs in adult β8 integrin KO mice show reduced self‐renewal, which is associated with abnormal SVZ ultrastructure. KO mice also develop size‐reduced olfactory bulbs, indicative of impaired NP growth and differentiation. Our various data for αvβ8 integrin are very similar to results for a regulator of NP self‐renewal, Bmi1. Indeed, our data mining efforts reveal that Bmi1 regulates αvβ8 integrin gene expression. Collectively, these studies identify essential roles for αvβ8 integrin in NP self‐renewal and differentiation in the CNS neurovascular niche.

The Catalog Archive Server Database Management System

The multiterabyte Sloan Digital Sky Survey's (SDSS's) catalog data is stored in a commercial relational database management system with SQL query access and a built-in query optimizer. The SDSS catalog archive server adds advanced data mining features to the DBMS to provide fast online access to the data.

A data mining-based subset selection for enhanced discrimination using iterative elimination of redundancy

The presence of redundant or irrelevant features in data mining may result in a mask of underlying patterns. Thus one often reduces the number of features by applying a feature selection technique. The objective of feature selection is to get a feature subset that has the best performance. This work proposes a new feature selection method using orthogonal filtering and nonlinear representation of data for an enhanced discrimination performance. An orthogonal filtering is implemented to remove unwanted variation of data. The proposed method adopts kernel principal component analysis, one of nonlinear kernel methods, to extract nonlinear characteristics of data and to reduce the dimensionality of data. The proposed feature selection method is based on the selection criterion of linear discriminant analysis in an environment of iterative backward feature elimination. The performance of the proposed method is compared with those of three different methods. The results showed that it outperforms the three methods. The use of filtering and a kernel method was shown to be a promising tool for an efficient feature selection.

Learning from the Data: Mining of Large High-Throughput Screening Databases

High-throughput screening (HTS) campaigns in pharmaceutical companies have accumulated a large amount of data for several million compounds over a couple of hundred assays. Despite the general awareness that rich information is hidden inside the vast amount of data, little has been reported for a systematic data mining method that can reliably extract relevant knowledge of interest for chemists and biologists. We developed a data mining approach based on an algorithm called ontology-based pattern identification (OPI) and applied it to our in-house HTS database. We identified nearly 1500 scaffold families with statistically significant structure-HTS activity profile relationships. Among them, dozens of scaffolds were characterized as leading to artifactual results stemming from the screening technology employed, such as assay format and/or readout. Four types of compound scaffolds can be characterized based on this data mining effort: tumor cytotoxic, general toxic, potential reporter gene assay artifact, and target family specific. The OPI-based data mining approach can reliably identify compounds that are not only structurally similar but also share statistically significant biological activity profiles. Statistical tests such as Kruskal-Wallis test and analysis of variance (ANOVA) can then be applied to the discovered scaffolds for effective assignment of relevant biological information. The scaffolds identified by our HTS data mining efforts are an invaluable resource for designing SAR-robust diversity libraries, generating in silico biological annotations of compounds on a scaffold basis, and providing novel target family specific scaffolds for focused compound library design.

Integration and dimensional modeling approaches for complex data warehousing

With the broad development of the World Wide Web, various kinds of heterogeneous data (including multimedia data) are now available to decision support tasks. A data warehousing approach is often adopted to prepare data for relevant analysis. Data integration and dimensional modeling indeed allow the creation of appropriate analysis contexts. However, the existing data warehousing tools are well-suited to classical, numerical data. They cannot handle complex data. In our approach, we adapt the three main phases of the data warehousing process to complex data. In this paper, we particularly focus on two main steps in complex data warehousing. The first step is data integration. We define a generic UML model that helps representing a wide range of complex data, including their possible semantic properties. Complex data are then stored in XML documents generated by a piece of software we designed. The second important phase we address is the preparation of data for dimensional modeling. We propose an approach that exploits data mining techniques to assist users in building relevant dimensional models.

Catalyzer: a novel tool for integrating, managing and publishing heterogeneous bioscience data

AbstractThe integrative ambitions of systems biology and neuroinformatics—to construct working models of the machinery of living cells and brains—will flounder unless researchers have access to the huge amounts of diverse experimental data being collected. However, the vast majority of bioscience research data that is gathered is never made available to other researchers, partly for the want of an adequate software for annotating experimental data, and partly for social reasons (researchers are rarely rewarded for publishing the actual data sets—just for journal articles summarizing findings).We have developed a novel software solution aimed at making it simpler for researchers to annotate and publish their research data. The first part of this solution is a desktop application, Catalyzer, which lets researchers structure their data at source, and complements existing ad hoc solutions in use in labs (including cryptic filenames, Word, Excel, paper lab books) while being simpler and more flexible than relational databases, which are too complex for most bioscience researchers to set up. The catalogs produced by Catalyzer are stored in XML with a user defined schema, which will simplify future data mining efforts across large numbers of distributed data sets. The approach can be summarized as ‘structure at source, integrate as required’, with the initial focus on enabling the researchers to structure their own research data; only then will other researchers be able to integrate across data sets. Copyright © 2006 John Wiley & Sons, Ltd.

Congruent Strategies for Carbohydrate Sequencing. 1. Mining Structural Details by MSn

This report is the first in a series of three focused on establishing congruent strategies for carbohydrate sequencing. The reports are divided into (i) analytical considerations that account for all aspects of small oligomer structure by MSn disassembly, (ii) database support using an ion fragment library and associated tools for high-throughput analysis, and (iii) a concluding algorithm for defining oligosaccharide topology from MSn disassembly pathways. The analytical contribution of this first report explores the limits of structural detail exposed by ion trap mass spectrometry with samples prepared as methyl derivatives and analyzed as metal ion adducts. This data mining effort focuses on correlating the fragments of small oligomers to stereospecific glycan structures, an outcome attributed to a combination of metal ion adduction and analyte conformation. Facile glycosidic cleavage introduces a point of lability (pyranosyl-1-ene) that upon collisional activation initiates subsequent ring fragmentation. Product masses and ion intensities vary with interresidue linkage, branching position, and monomer stereochemistry. Excessive fragmentation is the property of small oligomers where collisional energy within a smaller number of oscillators dissipates through extensive fragmentation. The procedures discussed in this report are unified into a singular strategy using an ion trap mass spectrometer with the sensitivity expected for electron multiplier detection. Although a small set of structures have been discussed, the basic principles considered are fully congruent, with ample opportunities for expansion.

Evaluating indirect and direct classification techniques for network intrusion detection

The network intrusion detection domain has seen increased research that exploit data mining and machine learning techniques and principles. Typically, multi-category classification models are built to classify network traffic instances either as normal or belonging to a specific attack category. While many existing works on data mining in intrusion detection have focused on applying direct classification methods, to our knowledge indirect classification techniques have not been investigated for intrusion detection. In contrast to indirect classification techniques, direct classification techniques generally extend associated binary classifiers to handle multi-category classification problems. An indirect classification technique decomposes (binarization) the original multi-category problem into multiple binary classification problems. The classification technique used to train the set of binary classification problems is called the {base} classifier. Subsequently, a combining strategy is used to merge the results of the binary classifiers. We investigate two binarization techniques and three combining strategies, yielding six indirect classification methods. This study presents a comprehensive comparative study of five direct classification methods with the thirty indirect classification models (six indirect classification models for each of the five base classifiers). To our knowledge, there are no existing works that evaluate as many indirect classification techniques and compare them with direct classification methods, particularly for network intrusion detection. A case study of the DARPA KDD-1999 offline intrusion detection project is used to evaluate the different techniques. It is empirically shown that certain indirect classification techniques yield better network intrusion detection models.

Intellectual Property Strategies in Security and Privacy

With the US government increasingly tending towards eavesdropping on voice/online conversations and investing in data-mining efforts such as the Total Information Awareness project, and with foreign laws such as those requiring ISPs to keep email records for extended periods of time and to facilitate decryption of records by authorities, an increasing number of companies will develop surveillance technologies to be sold to governments. Such technologies will include data-mining algorithms for data, voice, and video information, firewalls and intrusion detectors, encryption/decryption techniques, massive storage systems, etc. implemented in software and/or hardware. The US Government’s push to increase national security has created many applications for biometric devices which have traditionally been used in local and small-scale security settings. On the other hand, technologies emerge to counter “Big Brother” and ensure citizens’ privacy. Examples involve the National ID controversy, unbreakable encryption schemes, untraceable and self-destructing email/publications, virtual private networks (VPNs), anonymous internet browsing, and eavesdrop detectors which will enable citizens to live privately in and out of cyberspace. Consumer demand for these technologies will increase harmoniously with the government’s increasing surveillance of what citizens consider private. IDC reports that the total IP VPN market will explode from $5.4 billion in 2001 to nearly $14.7 billion in 2006 [1]. The biggest issue in this area concerns the bottleneck of information that various devices gather without an efficient means of analysis to transform the information into a meaningful interpretation. Entrepreneurs who can solve the informational logjam riddle by developing new applications for these technologies--and can secure the associated patents-will be richly rewarded. With the high demand for these technologies that will likely occur, the probability that competitors will want to practice the associated patents is high, thus adding extra value to these registered patents.

Assessment of Driver Fatigue, Distraction, and Performance in a Naturalistic Setting

The results of a study to characterize episodes of driver fatigue and drowsiness and to assess the impact of driver fatigue on driving performance are documented. This data-mining effort performed additional analyses on data collected in an earlier study by the Federal Motor Carrier Safety Administration of the effects of fatigue on drivers in local and short-haul operations. The primary objectives of the study were to investigate fatigue as a naturally occurring phenomenon by identifying and characterizing episodes of drowsiness during all periods of driving and to determine the operational or driving environment factors associated with drowsy driving. A total of 2,745 drowsy events were identified in approximately 900 total hours of naturalistic driving video data. Higher levels of fatigue were associated with younger and less experienced drivers. In addition, a strong and consistent relationship was found between drowsiness and time of day. Drowsiness was twice as likely to occur between 6:00 a.m. and 9:00 a.m., and approximately 30% of all observed incidences of drowsiness occurred within the first hour of the work shift. Insights about the relationship between driver fatigue and driver distraction and inattention are provided. This study presents an analytical framework for quantitatively assessing driver fatigue and drowsiness as a function of driver characteristics and the driving environment. It is hoped that the results will help to identify effective countermeasures for drowsy driving that will reduce the number of commercial-vehicle-related fatalities and injuries.

Transcription-Based Prediction of Response to IFNβ Using Supervised Computational Methods

Changes in cellular functions in response to drug therapy are mediated by specific transcriptional profiles resulting from the induction or repression in the activity of a number of genes, thereby modifying the preexisting gene activity pattern of the drug-targeted cell(s). Recombinant human interferon beta (rIFNβ) is routinely used to control exacerbations in multiple sclerosis patients with only partial success, mainly because of adverse effects and a relatively large proportion of nonresponders. We applied advanced data-mining and predictive modeling tools to a longitudinal 70-gene expression dataset generated by kinetic reverse-transcription PCR from 52 multiple sclerosis patients treated with rIFNβ to discover higher-order predictive patterns associated with treatment outcome and to define the molecular footprint that rIFNβ engraves on peripheral blood mononuclear cells. We identified nine sets of gene triplets whose expression, when tested before the initiation of therapy, can predict the response to interferon beta with up to 86% accuracy. In addition, time-series analysis revealed potential key players involved in a good or poor response to interferon beta. Statistical testing of a random outcome class and tolerance to noise was carried out to establish the robustness of the predictive models. Large-scale kinetic reverse-transcription PCR, coupled with advanced data-mining efforts, can effectively reveal preexisting and drug-induced gene expression signatures associated with therapeutic effects.

MetaRouter: bioinformatics for bioremediation

Bioremediation, the exploitation of biological catalysts (mostly microorganisms) for removing pollutants from the environment, requires the integration of huge amounts of data from different sources. We have developed MetaRouter, a system for maintaining heterogeneous information related to bioremediation in a framework that allows its query, administration and mining (application of methods for extracting new knowledge). MetaRouter is an application intended for laboratories working in biodegradation and bioremediation, which need to maintain and consult public and private data, linked internally and with external databases, and to extract new information from it. Among the data-mining features is a program included for locating biodegradative pathways for chemical compounds according to a given set of constraints and requirements. The integration of biodegradation information with the corresponding protein and genome data provides a suitable framework for studying the global properties of the bioremediation network. The system can be accessed and administrated through a web interface. The full-featured system (except administration facilities) is freely available at http://pdg.cnb.uam.es/MetaRouter. Additional material: http://www.pdg.cnb.uam.es/biodeg_net/MetaRouter.

Predictive models for protein crystallization

Crystallization of proteins is a nontrivial task, and despite the substantial efforts in robotic automation, crystallization screening is still largely based on trial-and-error sampling of a limited subset of suitable reagents and experimental parameters. Funding of high throughput crystallography pilot projects through the NIH Protein Structure Initiative provides the opportunity to collect crystallization data in a comprehensive and statistically valid form. Data mining and machine learning algorithms thus have the potential to deliver predictive models for protein crystallization. However, the underlying complex physical reality of crystallization, combined with a generally ill-defined and sparsely populated sampling space, and inconsistent scoring and annotation make the development of predictive models non-trivial. We discuss the conceptual problems, and review strengths and limitations of current approaches towards crystallization prediction, emphasizing the importance of comprehensive and valid sampling protocols. In view of limited overlap in techniques and sampling parameters between the publicly funded high throughput crystallography initiatives, exchange of information and standardization should be encouraged, aiming to effectively integrate data mining and machine learning efforts into a comprehensive predictive framework for protein crystallization. Similar experimental design and knowledge discovery strategies should be applied to valid analysis and prediction of protein expression, solubilization, and purification, as well as crystal handling and cryo-protection.

Geometric Modeling of Highways Using Global Positioning System Data and B -Spline Approximation

The proliferation and increased precision of global positioning system (GPS) devices enable the technology to be used in new and innovative applications. In the past few years, the Kansas Department of Transportation collected spatial data on a highway system using annual GPS surveys. This research deals with the problem of handling very large amounts of geographical data (latitude, longitude, elevation, and additional attributes) that are not entirely consistent. This paper describes the methodology for cleaning up this large amount of data, condensing it, and generating a B-spline approximation of the highway. This data mining effort results in two separate two-dimensional models of the highway, represented as B-spline curves. The highway approximation methodology is validated using a comparison between the B-spline approximation and the actual design data for the highway.

Getting a clearer picture: a business application of data mining

Businesses are beginning to realise that knowledge discovery, and data mining in particular, are becoming an essential part of day-to-day operations. Yet, fundamentally technical and statistical techniques are not usually part of the expertise of the user in a business environment. Techniques that do not rely so heavily on such specialist knowledge could prove more accessible, and thus, more rewarding. Furthermore, the problem often lies not only with the technology and tools, but with the data mining process itself. This paper explores the experience gained from a cooperative data mining effort between the authors and a health insurance company, in order to suggest ways of overcoming these challenges. Firstly, technique complexity is addressed through the use of visual link analysis. Secondly, the CRISP-DM methodology is used as a framework within which the problems experienced by this business and possible ways in which they can be overcome are described. Visual link analysis is fundamentally based on using visual perception in order to discover results and proved relatively easy to use and understand. Whilst it was found that experts can initially provide valuable assistance in helping the business users come to terms with the data mining process, it was also found that the users’ business knowledge is essential to the process. Assistance proved especially useful in the analytical understanding and preparation of data, as well as the modeling. Automation of more technical tasks was required. In spite of this assistance, a number of further challenges in the process of this company institutionalising data mining are highlighted.

Maximum-likelihood crystallization

The crystallization facility of the TB Structural Genomics Consortium, one of nine NIH-sponsored structural genomics pilot projects, employs a combinatorial random sampling technique in high-throughput crystallization screening. Although data are still sparse and a comprehensive analysis cannot be performed at this stage, preliminary results appear to validate the random-screening concept. A discussion of statistical crystallization data analysis aims to draw attention to the need for comprehensive and valid sampling protocols. In view of limited overlap in techniques and sampling parameters between the publicly funded high-throughput crystallography initiatives, exchange of information should be encouraged, aiming to effectively integrate data mining efforts into a comprehensive predictive framework for protein crystallization.

Exploiting data mining techniques for broadcasting data in mobile computing environments

Mobile computers can be equipped with wireless communication devices that enable users to access data services from any location. In wireless communication, the server-to-client (downlink) communication bandwidth is much higher than the client-to-server (uplink) communication bandwidth. This asymmetry makes the dissemination of data to client machines a desirable approach. However, dissemination of data by broadcasting may induce high access latency in case the number of broadcast data items is large. We propose two methods aiming to reduce client access latency of broadcast data. Our methods are based on analyzing the broadcast history (i.e., the chronological sequence of items that have been requested by clients) using data mining techniques. With the first method, the data items in the broadcast disk are organized in such a way that the items requested subsequently are placed close to each other. The second method focuses on improving the cache hit ratio to be able to decrease the access latency. It enables clients to prefetch the data from the broadcast disk based on the rules extracted from previous data request patterns. The proposed methods are implemented on a Web log to estimate their effectiveness. It is shown through performance experiments that the proposed rule-based methods are effective in improving the system performance in terms of the average latency as well as the cache hit ratio of mobile clients.