Application Of Data Mining Methods Research Articles

A combination of variable selection and data mining techniques for high-dimensional statistical modelling

Variable selection is fundamental to statistical modelling in diverse fields of sciences. This paper deals with the problem of high-dimensional statistical modelling through the analysis of seismological data in Greece acquired during the years 1962-2003. The dataset consists of 10,333 observations and 11 factors, used to detect possible risk factors of large earthquakes. In our study, different statistical variable selection techniques are applied, while data mining techniques enable us to discover associations, meaningful patterns and rules. The statistical methods employed in this work were the non-concave penalised likelihood methods, SCAD, LASSO and Hard, the generalised linear logistic regression and the best subset variable selection. The applied data mining methods were three decision trees algorithms, the classification and regression tree (C&RT), the chi-square automatic interaction detection (CHAID) and the C5.0 algorithm. The way of identifying the significant variables in large datasets along with the performance of used techniques are also discussed.

Applying Data Mining Methods to Structural Identification

The main purpose of this paper is to present the possibilities of applying data mining methods to the problem of structural identification of complex systems, focusing on the static structural system properties. In this paper we will define the main set of problems of structural identification and propose the way of their solution by means of data mining methods. A specific solution to selected problems is demonstrated on the model designed in the Matlab program. In the framework of experiment based on the results of simulation, we will identify the existence of internal relations in the system and analyze the degrees of their dependencies.

Application of data mining methods to efficient microwave active device modeling

In this work, the signal and noise behaviors of a microwave transistor within its operation domain (voltage drain to source–VDS, current of drain to source—IDS, frequency—f) are modeled by data mining techniques (DMT) without using any information on the microwave circuit theory. The device is modeled by a black box whose small signal (S) and noise parameters are evaluated through data mining techniques, based on the fitting of both of these parameters for multiple bias and configuration. It has been shown that DMT have a high potential of faithful and efficient device modeling. © 2012 Wiley Periodicals, Inc. Int J RF and Microwave CAE, 2013.

Analysing factors related to slipping, stumbling, and falling accidents at work: Application of data mining methods to Finnish occupational accidents and diseases statistics database

The utilisation of data mining methods has become common in many fields. In occupational accident analysis, however, these methods are still rarely exploited. This study applies methods of data mining (decision tree and association rules) to the Finnish national occupational accidents and diseases statistics database to analyse factors related to slipping, stumbling, and falling (SSF) accidents at work from 2006 to 2007. SSF accidents at work constitute a large proportion (22%) of all accidents at work in Finland. In addition, they are more likely to result in longer periods of incapacity for work than other workplace accidents. The most important factor influencing whether or not an accident at work is related to SSF is the specific physical activity of movement. In addition, the risk of SSF accidents at work seems to depend on the occupation and the age of the worker. The results were in line with previous research. Hence the application of data mining methods was considered successful. The results did not reveal anything unexpected though. Nevertheless, because of the capability to illustrate a large dataset and relationships between variables easily, data mining methods were seen as a useful supplementary method in analysing occupational accident data.

Computational methods for climate data

AbstractAs climate changes, regional responses may become more apparent; impacts often can become natural hazards, adversely affecting millions of people, on all continents and most nations. The coupling of hazards to climate change scenarios is a great challenge of climate change science. Nevertheless, it is extremely important to observe, simulate and ultimately understand this coupling, for the benefit of society and sustainability of the Earth's environment. Different applied mathematical techniques have been used to discern real effects of these changes and study long term trends. Moreover, those techniques can be applied in addressing the intensity and frequency of extreme events associated with climate change at regional scales and would be an important step in facing future extreme events associated with climate change. Computational methods include applying statistical data analysis, mesoscale and climate simulations while assisting modeling efforts with satellite based observations. Predictive analytics platforms would be very useful for assessing impacts of climate variability and change on the frequency and intensity of extreme events and how these extreme events can affect water and air quality issues globally. These tools are an innovative technology that applies data mining methods, predictive models, analysis and reporting to data, without the inherent limitations of current On Line Analytical Processing tools that suffer from cube rigidity, database explosion and dimensional constriction. Climate‐induced changes are complex and vary across a wide range of dimensions. An important part of predictive analytics platforms are their unlimited dimensionality and the segmentation of data by ‘physical’ or ‘performance’ characteristics. For example, a physical dimension might be the climate divisions in the state of California. A performance dimension could be the arithmetical, mathematical, or statistical segmentation of data based on its performance with regard to time. WIREs Comput Stat 2012 doi: 10.1002/wics.1213This article is categorized under: Applications of Computational Statistics > Computational Climate Change and Numerical Weather Forecasting Applications of Computational Statistics > Signal and Image Processing and Coding Data: Types and Structure > Time Series, Stochastic Processes, and Functional Data

Multi-instance genetic programming for predicting student performance in web based educational environments

A considerable amount of e-learning content is available via virtual learning environments. These platforms keep track of learners’ activities including the content viewed, assignments submission, time spent and quiz results, which all provide us with a unique opportunity to apply data mining methods. This paper presents an approach based on grammar guided genetic programming, G3P-MI, which classifies students in order to predict their final grade based on features extracted from logged data in a web based education system. Our proposal works with multiple instance learning, a relatively new learning framework that can eliminate the great number of missing values that appear when the problem is represented by traditional supervised learning. Experimental results are carried out on data sets with information about several courses and demonstrate that G3P-MI successfully achieves better accuracy and yields trade-off between such contradictory metrics as sensitivity and specificity compared to the most popular techniques of multiple instance learning. This method could be quite useful for early identification of students at risk, especially in very large classes, and allows the instructor to provide information about the most relevant activities to help students have a better chance to pass a course.

Towards an effective automatic query expansion process using an association rule mining approach

The steady growth in the size of textual document collections is a key progress-driver for modern information retrieval techniques whose effectiveness and efficiency are constantly challenged. Given a user query, the number of retrieved documents can be overwhelmingly large, hampering their efficient exploitation by the user. In addition, retaining only relevant documents in a query answer is of paramount importance for an effective meeting of the user needs. In this situation, the query expansion technique offers an interesting solution for obtaining a complete answer while preserving the quality of retained documents. This mainly relies on an accurate choice of the added terms to an initial query. Interestingly enough, query expansion takes advantage of large text volumes by extracting statistical information about index terms co-occurrences and using it to make user queries better fit the real information needs. In this respect, a promising track consists in the application of data mining methods to the extraction of dependencies between terms. In this paper, we present a novel approach for mining knowledge supporting query expansion that is based on association rules. The key feature of our approach is a better trade-off between the size of the mining result and the conveyed knowledge. Thus, our association rules mining method implements results from Galois connection theory and compact representations of rules sets in order to reduce the huge number of potentially useful associations. An experimental study has examined the application of our approach to some real collections, whereby automatic query expansion has been performed. The results of the study show a significant improvement in the performances of the information retrieval system, both in terms of recall and precision, as highlighted by the carried out significance testing using the Wilcoxon test.

Birth defects detection algorithm based on emerging patterns

The problem of birth defects is one of the most important public health problems in the world,and the application of data mining method to improve the diagnostic accuracy for birth defects is a hot medical research issue.The authors proposed two emerging patterns for birth defects feature extraction: the defection contrast to normal and the normal contrast to defection.The Birth Defects Detection based on Emerging Patterns(BDD-EP) algorithm was implemented through combining the proposed patterns with C4.5 decision tree.The extensive experimental results show that the detection accuracy of BDD-EP is as high as 90.1%,the F-measure of normal samples is 93.9%,and the F-measure of defect samples is 74.1%.Compared with other famous classical classification algorithms,BDD-EP algorithm can get better results.

Uncovering Actionable Knowledge in Corporate Data with Qualified Association Rules

Association rules mining is one of the most successfully applied data mining methods in today’s business settings (e.g. Amazon or Netflix recommendations to customers). Qualified association rules mining is an extension of the association rules data mining method, that uncovers previously unknown correlations that only manifest themselves under certain circumstances (e.g. on a particular day of the week), with the goal of improving action results, e.g. turning an underperforming campaign (spread too thin over the entire audience) into a highly targeted campaign that delivers results. Such correlations have not been easily reachable using standard data mining tools so far. This paper describes the method for straightforward discovery of qualified association rules and demonstrates the use of qualified association rules mining on an actual corporate data set. The data set is a subset of a corporate data warehouse for Sam’s Club, a division of Wal-Mart Stores, INC. The experiments described in this paper illustrate how qualified association rules supplement standard association rules data mining methods and provide additional information which can be used to better target corporate actions.

Web Usage Association Rule Mining System

Introduction Web usage log files generated on web servers contain huge amount of information suitable for applying data mining methods to discover potentially useful knowledge (Kosala & Blockeel, 2000; Wang, Li, & Zhang, 2005). Discovering web usage association rules is one of the popular data mining methods that can be applied on the web usage log data. The information contained in association rules can be used to learn about website visitor behaviour patterns, enhance website structure making it more effective for the visitors, or improve web marketing campaigns (Anand, Mulvenna, & Chavielier, 2004; Cooley, Mobasher, & Srivastava, 1997). Originally, association rule mining algorithms were applied to the analysis of transactional databases (Agrawal, Imielinski, & Swami, 1993; Brin, Motwani, Ullman, & Tsur, 1997). When evaluating the association rule interestingness, various measures can be used to help find the rules that give maximally useful information to the user (Geng & Hamilton, 2006; Tan, Kumar, & Srivastava, 2004). Some of the proposed association rule interestingness measures are all-confidence (Omiecinski, 2003), collective strength (Aggarwal & Yu, 1998), conviction and lift (Brin et al., 1997). While association rule finding algorithms are complete in that they find all rules that satisfy defined constraints, they often result in a huge set of rules that is difficult to exploit in order to find those rules that are truly interesting to the user (Liu, Hsu, & Ma, 1999). This problem is aggravated in association rule mining of the web usage log data (Huang, Cercone, & Aijun, 2002). Web usage data is specific and differs from the market basket data in the way that it contains a large number of tightly correlated items (web resources or web pages) due to the link structure of a website. Web pages that are tightly linked together often occur in the same visitor sessions, which is a reason that the generated set of association rules contains a huge number of so called hard association rules, which are not truly interesting to a user (Huang, 2007). There are freely available data mining software systems that can be used for discovering association rules in the web log usage data (Weka 3, n.d.). Weka is a great data mining tool with a wide range of features (Witten, Frank, & Hall, 2011). We used Weka in our previous research (Dimitrijevic & Bosnjak, 2010), but ran into several obstacles. We found that Weka did not contain an integrated set of tools to support all phases of web usage mining. A special software had to be used for web log preprocessing, such as WUM Prep scripts. Further, we had to develop our own special purpose converters of the web log preprocessed data into the Weka's ARFF file format before running web usage association rule discovery algorithm in Weka. Other limitations of the current Weka version are that it offers four association rule interestingess measures without option of combining them. Weka can be extended to include more interestingness measures, but it would involve additional work. There is also no support for automatic pruning of the discovered association rules in Weka using the methods we proposed (this was confirmed to us by the Weka architect in a forum). We concluded that it would give us more flexibility in our research and ability to apply the methods we propose for pruning association rules and for combining various interestingness measures if we used our own independent tool specialized for Web Usage Mining. We present the first version of this tool in this paper. We apply the system for the discovery of the association rules on a real life data set and present the results using various parameter values. We propose a method to alleviate the problem of over-generation of not truly interesting rules in web usage mining by eliminating the rules that contain directly linked pages. …

ANALYSIS OF STUDENTS’ STUDY ACTIVITIES IN VIRTUAL LEARNING ENVIRONMENTS USING DATA MINING METHODS / STUDENTŲ, BESIMOKANČIŲ VIRTUALAUS MOKYMO APLINKOJE, VEIKLOS ANALIZĖ TAIKANT DUOMENŲ GAVYBOS METODUS

This article deals with application of data mining methods’ to analysis of learners’ behaviour using the distance learning platform BlackBoard Vista (BlackBoard 2008). Before planning a distance learning course, instructors have to pay attention to the fact that there exist different study methods: some students start reading learning materials from the very beginning to the end, some students look at unclear topics only, some start with the discussions, etc. Therefore after analyzing the learning factors and identifying learner's style, it is possible to prepare individualized learning materials and to choose a proper way of course presentation. Such a way of study organization would improve the quality of studies and make it possible to reach better results. The research was performed by observing the behaviour and results achieved by 528 students in 15 distance learning courses and, using the clustering method, 3 learner's styles using virtual learning environments (VLE) have been identified and work methods proposed for students with regard to those learners’ styles. Besides, the research aims to find out the factors that influence final evaluations of students’. Santrauka Prieš planuodami rengti ir teikti nuotolinio mokymosi kursą, rengėjai turi atsižvelgti į tai, kad žmonės studijuoja skirtingais metodais: vieni pradeda skaityti pateiktą medžiagą iš eiles, kiti peržiūri tik nesuprantamas vietas, treti persikelia į virtualias diskusijas ir pan. Todėl, išanalizavus mokymosi veiksmus ir nustačius studento stilių, vėliau galima pateikti suasmenintą mokymosi medžiagą, parinkti geresnius kurso pateikimo metodus. Toks mokymo organizavimas pagerintų studijų kokybę ir leistų pasiekti geresnių rezultatų. Šiame straipsnyje nagrinėjamas duomenų gavybos metodu taikymas, analizuojant studentų elgseną, naudojantis virtualaus mokymo terpe BlackBoard Vista (BlackBoard 2008).

Data mining application on crash simulation data of occupant restraint system

This article presents an application of data mining method on finite element data and crashworthiness result data of an occupant restraint system. According to the characteristics of the CAE (Computer-Aided Engineering) data, a framework for data preparation is developed based on object-oriented programming concepts. Training sets are built from data recorded in 98 crash simulations that adhere to FMVSS208, the America occupant crash protection testing standard. Relationship between design parameters and system effectiveness is implied in these data sets. Decision tree using C4.5 algorithm and attribute selection method based on attribute’s estimated importance are introduced to perform data mining on the building of training sets. The result yielded by data mining endows us with a deeper insight into the interrelations between the key design parameters and the performance of the occupant restraint system in crash simulations. Finally, the learned rules are tested on the real crash simulation data sets. The result of the testing shows that these rules are proper, and can been used as a guidance for the design of the occupant restraint system.

Classification rule discovery for the aviation incidents resulted in fatality

Data mining methods have been successfully applied to different fields. Aviation industry is one of them. There is a large amount of knowledge and data accumulation in aviation industry. These data could be stored in the form of pilot reports, maintenance reports, incident reports or delay reports. This paper explains the data mining application on the incident reports of the Federal Aviation Administration (FAA) Accident/Incident Data System database, contains incident data records for all categories of civil aviation between the years of 2000 and 2006. In this study, we applied data mining methods on the incident reports. Moreover rough sets concept is used to reduce the attributes of data set. The purpose of this application is to find out the effective attributes in order to reduce the number of the fatality in the incidents. The categorization tools and decision trees are used to find the relations and rules about the incidents resulted in fatality. For this purpose data-mining analysis is conducted. As a result some rules about the fatality are obtained and also the parameters that affect the fatality in the incident have determined. The rules found are tested in terms of their accuracy and reliability, and these results are seen to be meaningful.

A novel similarity-measure for the analysis of genetic data in complex phenotypes

BackgroundRecent technological advances in DNA sequencing and genotyping have led to the accumulation of a remarkable quantity of data on genetic polymorphisms. However, the development of new statistical and computational tools for effective processing of these data has not been equally as fast. In particular, Machine Learning literature is limited to relatively few papers which are focused on the development and application of data mining methods for the analysis of genetic variability. On the other hand, these papers apply to genetic data procedures which had been developed for a different kind of analysis and do not take into account the peculiarities of population genetics. The aim of our study was to define a new similarity measure, specifically conceived for measuring the similarity between the genetic profiles of two groups of subjects (i.e., cases and controls) taking into account that genetic profiles are usually distributed in a population group according to the Hardy Weinberg equilibrium.ResultsWe set up a new kernel function consisting of a similarity measure between groups of subjects genotyped for numerous genetic loci. This measure weighs different genetic profiles according to the estimates of gene frequencies at Hardy-Weinberg equilibrium in the population. We named this function the "Hardy-Weinberg kernel".The effectiveness of the Hardy-Weinberg kernel was compared to the performance of the well established linear kernel. We found that the Hardy-Weinberg kernel significantly outperformed the linear kernel in a number of experiments where we used either simulated data or real data.ConclusionThe "Hardy-Weinberg kernel" reported here represents one of the first attempts at incorporating genetic knowledge into the definition of a kernel function designed for the analysis of genetic data. We show that the best performance of the "Hardy-Weinberg kernel" is observed when rare genotypes have different frequencies in cases and controls. The ability to capture the effect of rare genotypes on phenotypic traits might be a very important and useful feature, as most of the current statistical tools loose most of their statistical power when rare genotypes are involved in the susceptibility to the trait under study.

Using data mining to segment healthcare markets from patients' preference perspectives

This paper aims to provide an example of how to use data mining techniques to identify patient segments regarding preferences for healthcare attributes and their demographic characteristics. Data were derived from a number of individuals who received in-patient care at a health network in 2006. Data mining and conventional hierarchical clustering with average linkage and Pearson correlation procedures are employed and compared to show how each procedure best determines segmentation variables. Data mining tools identified three differentiable segments by means of cluster analysis. These three clusters have significantly different demographic profiles. The study reveals, when compared with traditional statistical methods, that data mining provides an efficient and effective tool for market segmentation. When there are numerous cluster variables involved, researchers and practitioners need to incorporate factor analysis for reducing variables to clearly and meaningfully understand clusters. Interests and applications in data mining are increasing in many businesses. However, this technology is seldom applied to healthcare customer experience management. The paper shows that efficient and effective application of data mining methods can aid the understanding of patient healthcare preferences.

Analyzing Solar Power Plant Performance Through Data Mining

This work is devoted to application of data mining methods for monitoring of state of a solar thermal plant. The methods discussed are illustrated by example of a study performed for the DISS direct steam generation facility at the Plataforma Solar de Almeria (Spain). In order to deal with the problems of large dimensionality and high correlation among the data the methods of latent variables, principal component analysis and partial least squares, were applied. Results showed that normal and abnormal states during plant operation could be identified.

Analysis of cognitive performance in schizophrenia patients and healthy individuals with unsupervised clustering models

Currently, assignment of cognitive test results to particular cognitive domains is guided by theoretical considerations and expert judgments which may vary. More objective means of classification may advance understanding of the relationships between test performance and the cognitive functions probed. We examined whether “atheoretical” analyses of cognitive test data can help identify potential hidden structures in cognitive performance. Novel data-mining methods which “let the data talk” without a priori theoretically bound constraints were used to analyze neuropsychological test results of 75 schizophrenia patients and 57 healthy individuals. The analyses were performed on the combined sample to maximize the “atheoretical” approach and allow it to reveal different structures of cognition in patients and controls. Analyses used unsupervised clustering methods, including hierarchical clustering, self-organizing maps (SOM), k-means and supermagnetic clustering (SPC). The model revealed two major clusters containing accuracy and reaction time measures respectively. The sensitivity (75% versus 52%) and specificity (95% versus 77% ) of these clusters for diagnosing schizophrenia differed. Downstream branching was influenced by stimulus domain. Predictions arising from this “atheoretical” model are supported by evidence from published studies. This preliminary study suggests that appropriate application of data-mining methods may contribute to investigation of cognitive functions.

Effective Data Mining for a Transportation Information System

This paper describes the application of data mining methods in the database of the DORIS transportation information system, currently used by the Prague Public Transit Company. The goal is to create knowledge about the behavior of objects within this information system. Data is analyzed partly with the help of descriptive statistical methods, and partly with the help of association rules, which may discover common combinations of attributes that occur most frequently within a given data set. Two types of quantifiers were used when creating the association rules; namely “founded implication” and “above average”. The results of the analysis are presented in the form of graphs and hypotheses.

Integration of OLAP and data mining for analysis of results from dependability evaluation experiments

This paper proposes the application of On-Line Analytical Processing (OLAP) and data mining approaches to analyse the large amount of raw data collected in fault injection campaigns and dependability benchmarking experiments. We use data warehousing technologies to store raw results from different experiments in a multidimensional structure where raw data can be analysed by means of OLAP tools. Moreover, we present an approach for identifying the key infrastructural factors determining the behaviour of computer systems in the presence of faults by the application of data mining methods on the data sets. Results obtained with the proposed techniques identified important factors impacting performance and dependability that could not have been revealed solely by the benchmark measures.

Mining data from intensive care patients

In this paper we describe the application of data mining methods for predicting the evolution of patients in an intensive care unit. We discuss the importance of such methods for health care and other application domains of engineering. We argue that this problem is an important but challenging one for the current state of the art data mining methods and explain what improvements on current methods would be useful. We present a promising study on a preliminary data set that demonstrates some of the possibilities in this area.