Ensemble Data Mining Methods, also known as Committee Methods or Model Combiners, are machine learning methods that leverage the power of multiple models to achieve better prediction accuracy than any of the individual models could on their own. The basic goal when designing an ensemble is the same as when establishing a committee of people: each member of the committee should be as competent as possible, but the members should be complementary to one another. If the members are not complementary, i.e., if they always agree, then the committee is unnecessary---any one member is sufficient. If the members are complementary, then when one or a few members make an error, the probability is high that the remaining members can correct this error. Research in ensemble methods has largely revolved around designing ensembles consisting of competent yet complementary models.

Ensemble Data Mining Methods, also known as Committee Methods or Model Combiners, are machine learning methods that leverage the power of multiple models to achieve better prediction accuracy than any of the individual models could on their own. The basic goal when designing an ensemble is the same as when establishing a committee of people: each member of the committee should be as competent as possible, but the members should be complementary to one another. If the members are not complementary, i.e., if they always agree, then the committee is unnecessary---any one member is sufficient. If the members are complementary, then when one or a few members make an error, the probability is high that the remaining members can correct this error. Research in ensemble methods has largely revolved around designing ensembles consisting of competent yet complementary models.

Due to increasing use of technology-enhanced educational assessment, data mining methods have been explored to analyse process data in log files from such assessment. However, most studies were limited to one data mining technique under one specific scenario. The current study demonstrates the usage of four frequently used supervised techniques, including Classification and Regression Trees (CART), gradient boosting, random forest, support vector machine (SVM), and two unsupervised methods, Self-organizing Map (SOM) and k-means, fitted to one assessment data. The USA sample (N = 426) from the 2012 Program for International Student Assessment (PISA) responding to problem-solving items is extracted to demonstrate the methods. After concrete feature generation and feature selection, classifier development procedures are implemented using the illustrated techniques. Results show satisfactory classification accuracy for all the techniques. Suggestions for the selection of classifiers are presented based on the research questions, the interpretability and the simplicity of the classifiers. Interpretations for the results from both supervised and unsupervised learning methods are provided.

This chapter presents theoretical and practical aspects associated to the implementation of a combined model-based/data-driven approach for failure prognostics based on particle filtering algorithms, in which the current esti- mate of the state PDF is used to determine the operating condition of the system and predict the progression of a fault indicator, given a dynamic state model and a set of process measurements. In this approach, the task of es- timating the current value of the fault indicator, as well as other important changing parameters in the environment, involves two basic steps: the predic- tion step, based on the process model, and an update step, which incorporates the new measurement into the a priori state estimate. This framework allows to estimate of the probability of failure at future time instants (RUL PDF) in real-time, providing information about time-to- failure (TTF) expectations, statistical confidence intervals, long-term predic- tions; using for this purpose empirical knowledge about critical conditions for the system (also referred to as the hazard zones). This information is of paramount significance for the improvement of the system reliability and cost-effective operation of critical assets, as it has been shown in a case study where feedback correction strategies (based on uncertainty measures) have been implemented to lengthen the RUL of a rotorcraft transmission system with propagating fatigue cracks on a critical component. Although the feed- back loop is implemented using simple linear relationships, it is helpful to provide a quick insight into the manner that the system reacts to changes on its input signals, in terms of its predicted RUL. The method is able to manage non-Gaussian pdf’s since it includes concepts such as nonlinear state estimation and confidence intervals in its formulation. Real data from a fault seeded test showed that the proposed framework was able to anticipate modifications on the system input to lengthen its RUL. Results of this test indicate that the method was able to successfully suggest the correction that the system required. In this sense, future work will be focused on the development and testing of similar strategies using different input-output uncertainty metrics.

We discuss a statistical framework that underlies envelope detection schemes as well as dynamical models based on Hidden Markov Models (HMM) that can encompass both discrete and continuous sensor measurements for use in Integrated System Health Management (ISHM) applications. The HMM allows for the rapid assimilation, analysis, and discovery of system anomalies. We motivate our work with a discussion of an aviation problem where the identification of anomalous sequences is essential for safety reasons. The data in this application are discrete and continuous sensor measurements and can be dealt with seamlessly using the methods described here to discover anomalous flights. We specifically treat the problem of discovering anomalous features in the time series that may be hidden from the sensor suite and compare those methods to standard envelope detection methods on test data designed to accentuate the differences between the two methods. Identification of these hidden anomalies is crucial to building stable, reusable, and cost-efficient systems. We also discuss a data mining framework for the analysis and discovery of anomalies in high-dimensional time series of sensor measurements that would be found in an ISHM system. We conclude with recommendations that describe the tradeoffs in building an integrated scalable platform for robust anomaly detection in ISHM applications.

datasets that we used for mining sequential patterns

This data set belongs to the paper "Video-to-Model: Unsupervised Trace Extraction from Videos for Process Discovery and Conformance Checking in Manual Assembly", submitted on March 24, 2020, to the 18th International Conference on Business Process Management (BPM).Abstract: Manual activities are often hidden deep down in discrete manufacturing processes. For the elicitation and optimization of process behavior, complete information about the execution of Manual activities are required. Thus, an approach is presented on how execution level information can be extracted from videos in manual assembly. The goal is the generation of a log that can be used in state-of-the-art process mining tools. The test bed for the system was lightweight and scalable consisting of an assembly workstation equipped with a single RGB camera recording only the hand movements of the worker from top. A neural network based real-time object classifier was trained to detect the worker’s hands. The hand detector delivers the input for an algorithm, which generates trajectories reflecting the movement paths of the hands. Those trajectories are automatically assigned to work steps using the position of material boxes on the assembly shelf as reference points and hierarchical clustering of similar behaviors with dynamic time warping. The system has been evaluated in a task-based study with ten participants in a laboratory, but under realistic conditions. The generated logs have been loaded into the process mining toolkit ProM to discover the underlying process model and to detect deviations from both, instructions and ground truth, using conformance checking. The results show that process mining delivers insights about the assembly process and the system’s precision.The data set contains the generated and the annotated logs based on the video material gathered during the user study. In addition, the petri nets from the process discovery and conformance checking conducted with ProM (http://www.promtools.org) and the reference nets modeled with Yasper (http://www.yasper.org/) are provided.

Missing data is an inevitable aspect of every empirical research. Researchers developed several techniques to handle missing data to avoid information loss and biases. Over the past 50 years, these methods have become more and more efficient and also more complex. Building on previous review studies, this paper aims to analyze what kind of missing data handling methods are used among various scientific disciplines. For the analysis, we used nearly 50.000 scientific articles that were published between 1999 and 2016. JSTOR provided the data in text format. Furthermore, we utilized a text-mining approach to extract the necessary information from our corpus. Our results show that the usage of advanced missing data handling methods such as Multiple Imputation or Full Information Maximum Likelihood estimation is steadily growing in the examination period. Additionally, simpler methods, like listwise and pairwise deletion, are still in widespread use.

Peer-to-Peer (P2P) networks are gaining increasing popularity in many distributed applications such as file-sharing, network storage, web caching, sear- ching and indexing of relevant documents and P2P network-threat analysis. Many of these applications require scalable analysis of data over a P2P network. This paper starts by offering a brief overview of distributed data mining applications and algorithms for P2P environments. Next it discusses some of the privacy concerns with P2P data mining and points out the problems of existing privacy-preserving multi-party data mining techniques. It further points out that most of the nice assumptions of these existing privacy preserving techniques fall apart in real-life applications of privacy-preserving distributed data mining (PPDM). The paper offers a more realistic formulation of the PPDM problem as a multi-party game and points out some recent results.

This research used data mining approaches to better understand factors affecting the formation of secondary organic aerosol (SOA). Although numerous laboratory and computational studies have been completed on SOA formation, it is still challenging to determine factors that most influence SOA formation. Experimental data were based on previous work described by Offenberg et al. (2017), where volume concentrations of SOA were measured in 139 laboratory experiments involving the oxidation of single hydrocarbons under different operating conditions. Three different data mining methods were used, including nearest neighbor, decision tree, and pattern mining. Both decision tree and pattern mining approaches identified similar chemical and experimental conditions that were important to SOA formation. Among these important factors included the number of methyl groups, the number of rings and the presence of dinitrogen pentoxide (N2O5). This dataset is associated with the following publication: Olson, D., J. Offenberg, M. Lewandowski, T. Kleindienst, K. Docherty, M. Jaoui, J.D. Krug, and T. Riedel. Data mining approaches to understanding the formation of secondary organic aerosol. ATMOSPHERIC ENVIRONMENT. Elsevier Science Ltd, New York, NY, USA, 252: 118345, (2021).

Technical notes and documentation on the common data model of the project CONCEPT-DM2.

This publication corresponds to the Common Data Model (CDM) specification of the CONCEPT-DM2 project for the implementation of a federated network analysis of the healthcare pathway of type 2 diabetes.

Aims of the CONCEPT-DM2 project:

General aim: To analyse chronic care effectiveness and efficiency of care pathways in diabetes, assuming the relevance of care pathways as independent factors of health outcomes using data from real life world (RWD) from five Spanish Regional Health Systems.

Main specific aims:

• To characterize the care pathways in patients with diabetes through the whole care system in terms of process indicators and pharmacologic recommendations
• To compare these observed care pathways with the theoretical clinical pathways derived from the clinical practice guidelines
• To assess if the adherence to clinical guidelines influence on important health outcomes, such as cardiovascular hospitalizations.
• To compare the traditional analytical methods with process mining methods in terms of modeling quality, prediction performance and information provided.

Study Design: It is a population-based retrospective observational study centered on all T2D patients diagnosed in five Regional Health Services within the Spanish National Health Service. We will include all the contacts of these patients with the health services using the electronic medical record systems including Primary Care data, Specialized Care data, Hospitalizations, Urgent Care data, Pharmacy Claims, and also other registers such as the mortality and the population register.

Cohort definition: All patients with code of Type 2 Diabetes in the clinical health records

• Inclusion criteria: patients that, at 01/01/2017 or during the follow-up from 01/01/2017 to 31/12/2022 had active health card (active TIS - tarjeta sanitaria activa) and code of type 2 diabetes (T2D, DM2 in spanish) in the clinical records of primary care (CIAP2 T90 in case of using CIAP code system)
• Exclusion criteria:
  • patients with no contact with the health system from 01/01/2017 to 31/12/2022
  • patients that had a T1D (DM1) code opened after the T2D code during the follow-up.
• Study period. From 01/01/2017 to 31/12/2022

Files included in this publication:

• Datamodel_CONCEPT_DM2_diagram.png
• Common data model specification (Datamodel_CONCEPT_DM2_v.0.1.0.xlsx)
• Synthetic datasets (Datamodel_CONCEPT_DM2_sample_data)
  • sample_data1_dm_patient.csv
  • sample_data2_dm_param.csv
  • sample_data3_dm_patient.csv
  • sample_data4_dm_param.csv
  • sample_data5_dm_patient.csv
  • sample_data6_dm_param.csv
  • sample_data7_dm_param.csv
  • sample_data8_dm_param.csv
• Datamodel_CONCEPT_DM2_explanation.pptx

Results of data mining models.

ObjectiveOver the past decades, many studies have used data mining technology to predict the 5-year survival rate of colorectal cancer, but there have been few reports that compared multiple data mining algorithms to the TNM classification of malignant tumors (TNM) staging system using a dataset in which the training and testing data were from different sources. Here we compared nine data mining algorithms to the TNM staging system for colorectal survival analysis. MethodsTwo different datasets were used: 1) the National Cancer Institute's Surveillance, Epidemiology, and End Results dataset; and 2) the dataset from a single Chinese institution. An optimization and prediction system based on nine data mining algorithms as well as two variable selection methods was implemented. The TNM staging system was based on the 7th edition of the American Joint Committee on Cancer TNM staging system. ResultsWhen the training and testing data were from the same sources, all algorithms had slight advantages over the TNM staging system in predictive accuracy. When the data were from different sources, only four algorithms (logistic regression, general regression neural network, Bayesian networks, and Naïve Bayes) had slight advantages over the TNM staging system. Also, there was no significant differences among all the algorithms (p>0.05). ConclusionsThe TNM staging system is simple and practical at present, and data mining methods are not accurate enough to replace the TNM staging system for colorectal cancer survival prediction. Furthermore, there were no significant differences in the predictive accuracy of all the algorithms when the data were from different sources. Building a larger dataset that includes more variables may be important for furthering predictive accuracy.

The purpose of data mining analysis is always to find patterns of the data using certain kind of techiques such as classification or regression. It is not always feasible to apply classification algorithms directly to dataset. Before doing any work on the data, the data has to be pre-processed and this process normally involves feature selection and dimensionality reduction. We tried to use clustering as a way to reduce the dimension of the data and create new features. Based on our project, after using clustering prior to classification, the performance has not improved much. The reason why it has not improved could be the features we selected to perform clustering are not well suited for it. Because of the nature of the data, classification tasks are going to provide more information to work with in terms of improving knowledge and overall performance metrics. From the dimensionality reduction perspective: It is different from Principle Component Analysis which guarantees finding the best linear transformation that reduces the number of dimensions with a minimum loss of information. Using clusters as a technique of reducing the data dimension will lose a lot of information since clustering techniques are based a metric of 'distance'. At high dimensions euclidean distance loses pretty much all meaning. Therefore using clustering as a "Reducing" dimensionality by mapping data points to cluster numbers is not always good since you may lose almost all the information. From the creating new features perspective: Clustering analysis creates labels based on the patterns of the data, it brings uncertainties into the data. By using clustering prior to classification, the decision on the number of clusters will highly affect the performance of the clustering, then affect the performance of classification. If the part of features we use clustering techniques on is very suited for it, it might increase the overall performance on classification. For example, if the features we use k-means on are numerical and the dimension is small, the overall classification performance may be better. We did not lock in the clustering outputs using a random_state in the effort to see if they were stable. Our assumption was that if the results vary highly from run to run which they definitely did, maybe the data just does not cluster well with the methods selected at all. Basically, the ramification we saw was that our results are not much better than random when applying clustering to the data preprocessing. Finally, it is important to ensure a feedback loop is in place to continuously collect the same data in the same format from which the models were created. This feedback loop can be used to measure the model real world effectiveness and also to continue to revise the models from time to time as things change.

This chapter presents theoretical and practical aspects associated to the implementation of a combined model-based/data-driven approach for failure prognostics based on particle filtering algorithms, in which the current esti- mate of the state PDF is used to determine the operating condition of the system and predict the progression of a fault indicator, given a dynamic state model and a set of process measurements. In this approach, the task of es- timating the current value of the fault indicator, as well as other important changing parameters in the environment, involves two basic steps: the predic- tion step, based on the process model, and an update step, which incorporates the new measurement into the a priori state estimate. This framework allows to estimate of the probability of failure at future time instants (RUL PDF) in real-time, providing information about time-to- failure (TTF) expectations, statistical confidence intervals, long-term predic- tions; using for this purpose empirical knowledge about critical conditions for the system (also referred to as the hazard zones). This information is of paramount significance for the improvement of the system reliability and cost-effective operation of critical assets, as it has been shown in a case study where feedback correction strategies (based on uncertainty measures) have been implemented to lengthen the RUL of a rotorcraft transmission system with propagating fatigue cracks on a critical component. Although the feed- back loop is implemented using simple linear relationships, it is helpful to provide a quick insight into the manner that the system reacts to changes on its input signals, in terms of its predicted RUL. The method is able to manage non-Gaussian pdf’s since it includes concepts such as nonlinear state estimation and confidence intervals in its formulation. Real data from a fault seeded test showed that the proposed framework was able to anticipate modifications on the system input to lengthen its RUL. Results of this test indicate that the method was able to successfully suggest the correction that the system required. In this sense, future work will be focused on the development and testing of similar strategies using different input-output uncertainty metrics.

Nearly two thirds of surveyed top managers of large companies operating in Russia viewed process mining as useful for purchasing, in 2021. Furthermore, over ** percent of respondents saw the technology's potential in improving the customer journey map and IT processes.

Data supporting the Master thesis "Monitoring von Open Data Praktiken - Herausforderungen beim Auffinden von Datenpublikationen am Beispiel der Publikationen von Forschenden der TU Dresden" (Monitoring open data practices - challenges in finding data publications using the example of publications by researchers at TU Dresden) - Katharina Zinke, Institut für Bibliotheks- und Informationswissenschaften, Humboldt-Universität Berlin, 2023

This ZIP-File contains the data the thesis is based on, interim exports of the results and the R script with all pre-processing, data merging and analyses carried out. The documentation of the additional, explorative analysis is also available. The actual PDFs and text files of the scientific papers used are not included as they are published open access.

The folder structure is shown below with the file names and a brief description of the contents of each file. For details concerning the analyses approach, please refer to the master's thesis (publication following soon).

## Data sources

Folder 01_SourceData/

- PLOS-Dataset_v2_Mar23.csv (PLOS-OSI dataset)

- ScopusSearch_ExportResults.csv (export of Scopus search results from Scopus)

- ScopusSearch_ExportResults.ris (export of Scopus search results from Scopus)

- Zotero_Export_ScopusSearch.csv (export of the file names and DOIs of the Scopus search results from Zotero)

## Automatic classification

Folder 02_AutomaticClassification/

- (NOT INCLUDED) PDFs folder (Folder for PDFs of all publications identified by the Scopus search, named AuthorLastName_Year_PublicationTitle_Title)

- (NOT INCLUDED) PDFs_to_text folder (Folder for all texts extracted from the PDFs by ODDPub, named AuthorLastName_Year_PublicationTitle_Title)

- PLOS_ScopusSearch_matched.csv (merge of the Scopus search results with the PLOS_OSI dataset for the files contained in both)

- oddpub_results_wDOIs.csv (results file of the ODDPub classification)

- PLOS_ODDPub.csv (merge of the results file of the ODDPub classification with the PLOS-OSI dataset for the publications contained in both)

## Manual coding

Folder 03_ManualCheck/

- CodeSheet_ManualCheck.txt (Code sheet with descriptions of the variables for manual coding)

- ManualCheck_2023-06-08.csv (Manual coding results file)

- PLOS_ODDPub_Manual.csv (Merge of the results file of the ODDPub and PLOS-OSI classification with the results file of the manual coding)

## Explorative analysis for the discoverability of open data

Folder04_FurtherAnalyses

Proof_of_of_Concept_Open_Data_Monitoring.pdf (Description of the explorative analysis of the discoverability of open data publications using the example of a researcher) - in German

## R-Script

Analyses_MA_OpenDataMonitoring.R (R-Script for preparing, merging and analyzing the data and for performing the ODDPub algorithm)

IntroductionHospitals have seen a rise in Medical Emergency Team (MET) reviews. We hypothesised that the commonest MET calls result in similar treatments. Our aim was to design a pre-emptive management algorithm that allowed direct institution of treatment to patients without having to wait for attendance of the MET team and to model its potential impact on MET call incidence and patient outcomes.MethodsData was extracted for all MET calls from the hospital database. Association rule data mining techniques were used to identify the most common combinations of MET call causes, outcomes and therapies.ResultsThere were 13,656 MET calls during the 34-month study period in 7936 patients. The most common MET call was for hypotension [31%, (2459/7936)]. These MET calls were strongly associated with the immediate administration of intra-venous fluid (70% [1714/2459] v 13% [739/5477] p

Dataset

This dataset was created by Andre Guss

Contents

This research aimed at the case of customers’ default payments in Taiwan and compares the predictive accuracy of probability of default among six data mining methods. From the perspective of risk management, the result of predictive accuracy of the estimated probability of default will be more valuable than the binary result of classification - credible or not credible clients. Because the real probability of default is unknown, this study presented the novel “Sorting Smoothing Method†to estimate the real probability of default. With the real probability of default as the response variable (Y), and the predictive probability of default as the independent variable (X), the simple linear regression result (Y = A + BX) shows that the forecasting model produced by artificial neural network has the highest coefficient of determination; its regression intercept (A) is close to zero, and regression coefficient (B) to one. Therefore, among the six data mining techniques, artificial neural network is the only one that can accurately estimate the real probability of default.

Relevant Papers: Yeh, I. C., & Lien, C. H. (2009). The comparisons of data mining techniques for the predictive accuracy of probability of default of credit card clients. Expert Systems with Applications, 36(2), 2473-2480.