Search

With the advent of large-scale datasets, visualization has become one of the most useful concepts of data mining, that assists researchers to identify data patterns without manually scanning each of the entry sets. To perform this task, researchers have proposed a wide variety of visualization techniques, that include, Online Analytical Processing (OLAP), different chart types, map types, different graph representation types, etc. But each of these methods vary in terms of their applicability, scalability, and data representation characteristics. Due to which it is difficult for researchers to identify optimum techniques for their application-specific use cases. Moreover, these models showcase variations in terms of their data handling & analysis capabilities, due to which researchers are needed to test & validate different visualization techniques for their deployments. These issues affect selection capabilities of these visualization models, which increases cost & time to market for the underlying product. To overcome these issues, a comparison of these models in terms of their context-specific nuances, application-specific advantages, deployment-specific limitations, and functional future scopes are discussed in this text. Upon referring this discussion, researchers will be able to identify optimum models for their application-specific use cases. This text also compares the models in terms of their performance metrics including delay, complexity, usability, and scalability values, which will further assist readers to identify performance-specific models for different deployments. Thus, this text can be used as a reference model for comparing different visualization tools & techniques for multiple application scenarios. [ABSTRACT FROM AUTHOR]

The performance of query processing over OLAP (Online Analytical Processing) model is decreased due to higher query access time for huge multidimensional data. Therefore, the clustering is introduced to improve the OLAP model efficiency by getting quick query processing because of dividing the large data into various clusters. The K-Means is a famous technique of clustering the data into groups to solve various clusters. The K-Means is a famous technique of clustering the data into groups to solve various real life issues. However, K-Means has some drawbacks like sensitivity to primary centroid assortment in cluster and local optimum convergence. Hence, a KMeans-Salp Swarm Optimization based Clustering (K-SSOC) is implemented to improve the performance of K-Means by providing optimal clustering over huge OLAP multidimensional data. The outcomes are obtained on MATLAB 2019a environment based on the parameter purity index, standard deviation, F-measure, intra-cluster distance and running time complexity over 1000 iterations. The results illustrate the superior performance of K-SSOC against K-Means, ACO and PSO over total six multidimensional datasets based on parameters. [ABSTRACT FROM AUTHOR]

This article reviews MS-Power BI Desktop that empowers managers to see and understand their data. By drag and drop, it is possible to discover trends and outliers from interactive dashboards, including combining data from multiple sources (from spreadsheets and data warehouses to big cloud data) at a glance. If asking questions is a key role of any manager (either using a browser, a tablet or a phone), business intelligence (BI) is the tool that allows the change managerial perspective of the problem and reveal deeper meaning of the data. Inevitable, basic technical jargons such as data mining or on-line analytical processing (OLAP) are introduced too. Three practical cases are disclosed while its layouts are discussed on a managerial and academic perspective. [ABSTRACT FROM AUTHOR]

The article describes a methodological guide prepared for the development of a Data Warehousing Process (DWP) in a Chilean University. The goal of this DWP, is to generate indicators of academic productivity. For this, a methodological guide is elaborated, integrating diverse approaches and techniques, such as: specification of information requirements, relational modeling, combined development model of Kimball and Hefesto proposals, ETL process (Extraction, Transformation and Load) with a validation phase of indicators, and integrated and interactive visualizations, for the multidimensional analysis of the indicators with the use of dashboards. In this process, three elements incorporated that, in the opinion of the development team, favored the success and effectiveness of the development of the solution. Firstly, how to specify Key Performance Indicators (KPI) in detail, through the development of a template based on those used to specify information requirements in a software process. Second, to explain the incorporation of a validation phase of the KPI obtained in the ETL process, based on Structured Query Language queries (SQL), and the comparison with the data of the operational systems. Thirdly, integrated visualizations configured with an On-Line Analytical Processing (OLAP) tool, which interactively display the related indicators for the same activity under analysis. Because of this data warehousing process, you get a Business Intelligence (BI) platform based on a DataMart (DM) model with two schemes, one star and another snowflake. The first contains the indicators of teacher productivity, and the second, those of scientific productivity, which satisfy the specifications of the KPI, and defined in agreement with the end users. The methodological guide allowed, on the one hand, the complete development of the process in an appropriate time, and on the other hand, to ensure the quality of the indicators generated through an adequate validation phase. [ABSTRACT FROM AUTHOR]

Among various socio-pathological events, the crime is perceived as the most serious issue. Understanding of factors influencing the intensity, structure and dynamic of crime is essential for appropriate targeting of preventive efforts. Due to the complexity of crime, it is necessary to integrate various sources of data which are usually complicated by differences in spatial referencing, scale, temporal referencing, and data semantic. The study demonstrates data integration based on multidimensional modelling using Online Analytical Processing (OLAP) for analysis. Data from selected sources (e.g. population, crime, dwelling) were aggregated into 1 km grid with an appropriate temporal interval, transformed into relative indicators describing the local demographic and environmental features, and intensities of selected types of crime. The analysis includes evaluation of pairwise correlation and regression analysis (using backward method). The results in the Ostrava pilot area show the highest explanation of variability in the case of thefts (R2 0.60; important descriptors are several other types of crime but also higher share of retired population and unemployed with basic education), burglaries (R2 0.54, share of young population, long-term unemployed), violent crime and property offences (both R2 0.5). To the opposite, the overall index of crime can be explained by the set of independent factors from only 11%. It supports the idea of the selective influence of explored factors which differently affect particular types of crime. It was confirmed that the spatial distribution of selected types of crime intensity depends on the demographic features. The findings emphasize the role of social factors in crime preventive efforts and the necessity to utilize operative registers of public administration for data integration and monitoring of the current state of social conditions. [ABSTRACT FROM AUTHOR]

Online Analytical Processing (OLAP) is a commonly used approach to deliver business intelligence. This is generally accomplished in an environment which involves aggregations, drill-downs and slicing/dicing of data along multiple dimensions. The inherent complexity of building cubes for the purpose of analytics may be an overkill for applications which do not require multi-layered drill downs, but still need multi-dimensional analysis and reporting. We demonstrate achieving this goal by using a combination of web development and business intelligence tools, but bypassing OLAP cubes. An application to monitor student progress after enrollment in an academic program is used as an example. By eliminating the middle layer, performance is enhanced significantly and creates a low-maintenance environment. Rendering reports from a business intelligence tool stack, in combination with web forms for data integration, results in rapid application development and creates a stable, user-friendly environment which requires minimal operational support from the IT department. [ABSTRACT FROM AUTHOR]

Modern learning institutions, in particular those that carry on distance learning programmes and those that use advanced e- Learning platforms for the teaching activities they are conducting, are gathering more and more data in electronic format that can be analysed in order to obtain valuable information that can be applied to optimize their own activities. As long as the data volumes gathered are not too large, analysing it is not very difficult and most businesses that carry on learning activities can do such analysis without implying specialized tools. However, the benefits of analysing small volumes of data can be limited and for this reason most institutions do not do such analysis. On the other hand, when the data volumes recorded become larger and larger, there are both advantages and disadvantages. The main advantage that has to be mentioned is that the potential benefits become more and more important. The disadvantages that have also to be taken into account are that specialized computational tools and techniques are needed in order to be able to extract useful knowledge from the data available. Business intelligence type tools are based manly on the modern data analysis techniques that can be applied on large volumes of data. These type of tools can be applied by any business to extract useful knowledge from the data available and to improve their performances and gain an advantage in the competitive environment. The present paper aims to explore the possibilities of applying business intelligence type tools on the data available to learning institutions and find ways to enhance their performances in an environment that is more and more competitive. [ABSTRACT FROM AUTHOR]

The developed multidimensional data mining algorithm integrates logical and statistical computing. It provides interactive and exploratory environment for mining spatial and temporal features of time series datasets. The algorithm is applied on Crop production database which stores spatial time series on crop production in EU-Balkan states. The paper discusses the mined information on crop production quantity and area harvested in Bulgaria at national and regional level. The data mining algorithm and software can be used to analyze spatial time series when studying the dynamics of social systems, economical activities and natural resources exploitation. [ABSTRACT FROM AUTHOR]

Recently Model-based test case generation is attracting more attention of researchers in software development. This is due to the fact that this leads to the early detection of faults reducing software development time and cost. UML 2.0 activity diagram is one of the behavioral models used for test case generation. In this paper we propose an approach based on graph transformation to generate test cases from UML 2.0 activity diagrams. To this end, we propose two graph grammars. The first one transforms a UML activity diagram to an intermediate representation while the second one generates test cases from the intermediate representation. An example illustrates the approach. [ABSTRACT FROM AUTHOR]

The companies that apply the concept of business intelligence in their business in some of the following ways were included in this study: at the level of the entire system or specific strategic business units of companies (e.g. marketing department, research development, finance and accounting, manufacturing, commercial, procurement etc.), apply business intelligence only in certain business processes or projects, used in business from the technology and platform for data warehouse, data mining, OLAP tools, using advanced analytical techniques of simulation and visualization applications. Variables examined were categorized into four groups: business intelligence, supply chain management, information visibility and integration. Factor analysis was used to facilitate the connection of these groups of variables, i.e. reduction of number of variables. Then, we tested the correlation between the newly formed variables. There was a significant statistical correlation between business intelligence, supply chain management, information visibility and integration among the partners in the production chain. ANOVA was conducted to compare differences in the mean values of variables in relation to the activity, size and legal form. This paper will analyze the relationship between business intelligence and supply chain management. [ABSTRACT FROM AUTHOR]

OLAP (On Line Analytical Processing) is a set of techniques and operators to facilitate the data analysis usually stored in a data warehouse. In this paper, we extend the functionality of an OLAP operator known as Map Cube with the definition and incorporation of a function that allows the formulation of spatio-temporal queries. For example, consider a data warehouse about crimes that includes data about the places where the crimes were committed. Suppose we want to find and visualize the trajectory (a trajectory is just the path that an object follows through space as a function of time) of the crimes of a suspect beginning with his oldest crime and ending with his most recent one. In order to meet this requirement, we extend the Map Cube operator. [ABSTRACT FROM AUTHOR]

Allowing weights w∈]-1;1]\{0}, pseudo-convex mixtures increase usual mixtures flexibility. The finite mixtures with negative components are investigated for distribution families closed under minimization. The main purpose is to define these mixtures and to study their properties. [ABSTRACT FROM AUTHOR]

The increasing availability of mobility data and the awareness of its importance and value have been motivating many researchers to the development of models and tools for analyzing movement data. This paper presents a brief survey of significant research works about modeling, processing and visualization of data about moving objects. We identified some key research fields that will provide better features for online analysis of movement data. As result of the literature review, we suggest a generic multi-layer architecture for the development of an online analysis processing software tool, which will be used for the definition of the future work of our team. [ABSTRACT FROM AUTHOR]

Every business is dynamic in nature and is affected by various external and internal factors. These factors include external market conditions, competitors, internal restructuring and re-alignment, operational optimization and paradigm shifts in the business itself. New regulations and restrictions, in combination with the above factors, contribute to the constant evolutionary nature of compelling, business-critical information; the kind of information that an organization needs to sustain and thrive. Business intelligence ("BI") is broad term that encapsulates the process of gathering information pertaining to a business and the market it functions in. This information when collated and analyzed in the right manner, can provide vital insights into the business and can be a tool to improve efficiency, reduce costs, reduce time lags and bring many positive changes. A business intelligence application helps to achieve precisely that. Successful organizations maximize the use of their data assets through business intelligence technology. The first data warehousing and decision support tools introduced companies to the power and benefits of accessing and analyzing their corporate data. Business users at every level found new, more sophisticated ways to analyze and report on the information mined from their vast data warehouses. Choosing a Business Intelligence offering is an important decision for an enterprise, one that will have a significant impact throughout the enterprise. The choice of a BI offering will affect people up and down the chain of command (senior management, analysts, and line managers) and across functional areas (sales, finance, and operations). It will affect business users, application developers, and IT professionals. BI applications include the activities of decision support systems (DSS), query and reporting, online analytical processing (OLAP), statistical analysis, forecasting, and data mining. Another way of phrasing this is that BI applications take data that is generated by the operations of an enterprise and translate that data into relevant and useful information for consumption by people throughout the enterprise. [ABSTRACT FROM AUTHOR]

Learning strategy in an intelligent learning system is generally elaborated on the basis of assessment of the following factors: learner's time for reaction, content of the learning object, amount of learning material in a learning object, learning object specification, e-learning medium and performance control. Current work proposes architecture for dynamic learning strategy design by implementing multidimensional analysis model of learning factors. The analysis model concerns on-line analytical processing (OLAP) of learner's data structured as multidimensional cube. Main components of the architecture are analysis agent for performing the OLAP operations on learner data cube, adaptation generator and knowledge selection agent for performing adaptive navigation in the learning object repository. The output of the analysis agent is involved in dynamic elaboration of learning strategy that fits best to learners profile and behavior. As a result an adaptive learning path for individual learner and for learner groups is generated. [ABSTRACT FROM AUTHOR]

Over the past two decades, literature on digital divide has largely focused on the important role of First-Level Digital Divide. Meanwhile, digital divide has been narrowly defined as the gap in Internet access between the information rich and the information poor. Such dichotomous classification provides a poor road map for technology use. In fact, studies have indicated that, as the material access divide lessened over the years, the digital divide did not diminish accordingly. This paper intended to look beyond accessibility and explored the difference of digital skills among users. 300 information equipments were provided for users in rural area. Data mining was conducted and digital skills of the users were analyzed. The results of content analysis suggested that the Second-Level Digital Divide in rural Taiwan is not significant. The largest disparity was found in the ability to use e-mail among users. In addition, gender difference in accessing the provided information equipments were suggested for future studies. [ABSTRACT FROM AUTHOR]

Most of the time, users experiment difficulties when interacting with databases in order to obtain valuable information from them. On one hand, there is an existing problem concerning the representation and understanding of data. On the other hand, there is usually a lack of effectiveness and easy-of-use in successfully accessing the right information. Additionally, the need for interactive visual tools to query data is mandatory to avoid the dependency on programming languages. To address these drawbacks, we introduce a visual web-based application that (a) infers Data Warehouse schemes, (b) visually represents the database structure to facilitate the interaction and creation of queries, and (c) easily presents the resulting data in real time. Our system is intended for end-users having basic knowledge of database query but being unskilled in query-based programming, and it is an attempt to reduce the learning burden in a gradual way. [ABSTRACT FROM AUTHOR]

Clustering has played a very important role in data mining. In this paper, a fast incremental clustering algorithm is proposed by changing the radius threshold value dynamically. The algorithm restricts the number of the final clusters and reads the original dataset only once. At the same time an inter-cluster dissimilarity measure taking into account the frequency information of the attribute values is introduced. It can be used for the categorical data. The experimental results on the mushroom dataset show that the proposed algorithm is feasible and effective. It can be used for the large-scale data set. [ABSTRACT FROM AUTHOR]

This paper describes the application of online analytical processing (OLAP) in the field of data mining. We discuss knowledge discovery in enterprise databases as well as data warehouses. We then present an enterprise decision modelling application. This internet application stores requests, mistakes and defects that are notified by customers for software modules. Data are retrieved from Oracle database by using Microsoft Excel. Excel's interactive visual interface helps to build data cube and get graphics for executive managers. This OLAP application is applicable to any ecommerce or supply-chain applications which accumulate huge data sets in order to have the potential to yield more valuable information. [ABSTRACT FROM AUTHOR]

We consider the problem of online mining of association rules in databases containing large numbers of transactions - and e specially in market-basket data. We focus primarily in the use of a proper data structure that would allow us to handle large numbers of data for online mining of association rules, and also in presenting and giving answers to new types of online queries. We borrow techniques used in information retrieval, like inverted file indexing and inverted file compression. Thus, the online mining is accomplished by preprocessing the data once and creating an index by using an inverted file. Next, if this inverted file is too large to be held in main memory as is, we compress it. The inverted file is now ready to instantaneously give answers to various types of queries. We present the traditional kinds of queries to which our approach is capable of answering as well a s we present some new kinds of queries that are supported. Finally we present how our approach could be used like a classic algorithm for finding large itemsets. [ABSTRACT FROM AUTHOR]

VERITAS is an array of four 12 m diameter Atmospheric Cherenkov Telescopes located in southern Arizona operating since January 2007. In this contribution we discuss the various software systems needed to operate the array including the data acquisition, online analysis, real time hardware monitoring and offline analysis. © 2007 American Institute of Physics [ABSTRACT FROM AUTHOR]

In my talk, I will share with you our experience in applying and deploying our source code mining technology in industry. In particular, the most valuable lesson we have learned is that sometimes there is a bigger problem in the real world that can really benefit from our technology but unfortunately we do not know about it until we closely work with industry. In 2004, motivated from some previous research work that pointed out copy-pasting as a major reason for majority of the bugs in device driver code, my students and I applied data mining technology (specifically frequent subsequence mining algorithms) in identifying copy-pasted code and also detecting forget-to-change bugs introduced during copy-pasting. The benefit of using data mining is that it is highly scalable (20 minutes for 4-5 millions lines of code), and can tolerate changes such as statement insertion/deletion/modification as well as variable name changes. When we released our tool called CP-Miner to the open source community, it attracted some inquiries from industry. These inquiries have motivated us to start a company to commercialize our tools. During the commercialization process, our customers have taught us that our technology can be applied to solve a major headache faced by many embedded system vendors such as storage companies, network devices, telecommunications, handhelds, electronics, etc. These companies typically have to maintain tens and even hundreds of active branches of similar software, one for slightly different devices or with different customization. These branches are 60-70% similar and are being developed in parallel after split (and are usually never merged back together). So when developers decide to fix a bug or security hole in one branch, it is usually a big challenge for them to check what other branches a similar fix should be also applied. So many costly incidents have happened in the field due to some known bugs (a bug that has already been diagnosed and fixed in some other branches). Since these branches have diverged over the years, many code are similar but are not exactly the same. Also, file names can change, etc. Therefore, it is hard to rely on source control systems such as ClearCase, Subversion to keep track their differences. Being pushed by many customers who suffer this pain, we applied our technology to this problem and build a tool called PatchMiner. Currently PatchMiner has been deployed and widely used in several large companies (more information can be found at http://www.patterninsight.com/). It is a very interesting journey. To me personally, I learned that sometimes what I (as an academic) feel as a solution to a major problem may be only a nice-to-have in the real world; and it really requires close interaction with industry to understand their painpoints. [ABSTRACT FROM AUTHOR]