Search

Purpose: To gain an understanding of the heterogeneous group of type 2 diabetes (T2D) patients, we aimed to identify patients with the homogenous long-term HbA1c trajectories and to predict the trajectory membership for each patient using explainable machine learning methods and different clinical-, treatment-, and socio-economic-related predictors. Patients and Methods: Electronic health records data covering primary and specialized healthcare on 9631 patients having T2D diagnosis were extracted from the North Karelia region, Finland. Six-year HbA1c trajectories were examined with growth mixture models. Linear discriminant analysis and neural networks were applied to predict the trajectory membership individually. Results: Three HbA1c trajectories were distinguished over six years: “stable, adequate” (86.5%), “improving, but inadequate” (7.3%), and “fluctuating, inadequate” (6.2%) glycemic control. Prior glucose levels, duration of T2D, use of insulin only, use of insulin together with some oral antidiabetic medications, and use of only metformin were the most important predictors for the long-term treatment balance. The prediction model had a balanced accuracy of 85% and a receiving operating characteristic area under the curve of 91%, indicating high performance. Moreover, the results based on SHAP (Shapley additive explanations) values show that it is possible to explain the outcomes of machine learning methods at the population and individual levels. Conclusion: Heterogeneity in long-term glycemic control can be predicted with confidence by utilizing information from previous HbA1c levels, fasting plasma glucose, duration of T2D, and use of antidiabetic medications. In future, the expected development of HbA1c could be predicted based on the patient’s unique risk factors offering a practical tool for clinicians to support treatment planning.

Artificial intelligence (AI) is increasingly being used to improve patient care and management. In this paper, we propose explainable AI (XAI) models for predicting severe hypoglycemia (SH) and diabetic ketoacidosis (DKA) episodes in adults with type 1 diabetes (T1D) and relapses in adults with relapsing-remitting multiple sclerosis (RRMS). We follow a three-step process in this study: (1) develop baseline machine learning (ML) models, (2) improve the models using ReliefF feature selection technique, and develop sex-stratified models, (3) explain the models and their results using SHapley Additive exPlanations (SHAP). We built six ML models (XGBoost, LightGBM, CatBoost, AdaBoost, random forest, and linear regression) for all scenarios. Applying the ReliefF feature selection led to improved model performance in predicting all outcomes compared to the baseline models. Additionally, sex-stratified models further improved the prediction of SH episodes and relapses. The F1 scores for predicting SH episodes in male and female patients were 84.07% and 84.95%, respectively, and the DKA prediction model achieved an F1 score of 78.67%. The proposed relapse prediction models outperformed existing models with F1 scores of 84.55% (males) and 76.11% (females), and ROCs of 70.26% (males) and 69.05% (females). Our results highlight the importance of considering sex differences, socioeconomic factors, and physical and mental health in medical outcome prediction. Boosting ML algorithms were found to be effective in detecting SH and DKA in T1D patients and relapses in RRMS patients compared to conventional tree-based ML and statistical models.

Clinical data analysis could lead to breakthroughs. However, clinical data contain sensitive information about participants that could be utilized for unethical activities, such as blackmailing, identity theft, mass surveillance, or social engineering. Data anonymization is a standard step during data collection, before sharing, to overcome the risk of disclosure. However, conventional data anonymization techniques are not foolproof and also hinder the opportunity for personalized evaluations. Much research has been done for synthetic data generation using generative adversarial networks and many other machine learning methods; however, these methods are either not free to use or are limited in capacity. This study evaluates the performance of an emerging tool named synthpop, an R package producing synthetic data as an alternative approach for data anonymization. This paper establishes data standards derived from the original data set based on the utilities and quality of information and measures variations in the synthetic data set to evaluate the performance of the data synthesis process. The methods to assess the utility of the synthetic data set can be broadly divided into two approaches: general utility and specific utility. General utility assesses whether synthetic data have overall similarities in the statistical properties and multivariate relationships with the original data set. Simultaneously, the specific utility assesses the similarity of a fitted model’s performance on the synthetic data to its performance on the original data. The quality of information is assessed by comparing variations in entropy bits and mutual information to response variables within the original and synthetic data sets. The study reveals that synthetic data succeeded at all utility tests with a statistically non-significant difference and not only preserved the utilities but also preserved the complexity of the original data set according to the data standard establish

In this paper we present procedures for processing raw data collected with moving vehicles and for fusing this data with digital map data. The goal is to have a better understanding of the city traffic via quantitative research on collected taxi data in relation to digital map properties. Map attributes are provided by Digiroad, which is a database of Finnish road and street network. We define methods to clean up data that has been collected with taxis equipped with on-board vehicle tracking devices from real customer service situations. Consequently, the driving behavior may be inconsistent and sensor data can be limited and contain errors. We explain procedures of preparing data; filtering the most obvious errors from the data set, map-matching moving object data, and fetching map attributes along the routes of the moving vehicles. The fetched properties, as well as other measurement data, are used for deriving statistics and illustrations to study driving behavior in downtown Oulu, Finland.

Industry is increasingly moving towards data-driven business. With an abundance of data, the problem is no longer how to get access to data but how to extract the most value from it. The extracted knowledge helps to control industrial processes efficiently and automatically. The quality of products can be improved by finding root causes behind poor quality, hence improving the yield and competitiveness of the whole plant. This thesis helps to understand the benefit of using statistical machine learning methods to improve manufacturing processes. With the methods presented, the future can be predicted based on historical data and data-driven decision support for the processes can be offered. The application area of this work is the steel industry. This work gives step-by-step advice for successfully implementing AI applications in the industry. In addition, methods for finding root causes behind poor quality are presented to improve the process and the quality of the products. Since data collected under challenging measurement conditions in the industry is never flawless, a quality model which can utilize incomplete data is also developed. Machine learning methods are used in this work to process data and to develop data-driven quality models, which help to predict the desired quality characteristics and hence support the decision-making of the workers managing the process. Both supervised and semi-supervised machine learning methods are used. In addition, explainable machine learning is used to increase the transparency of the models. With the help of these methods, predictive quality models are developed for four different datasets consisting of measurements from steel manufacturing processes. The developed models have been implemented as a smart manufacturing tool, which enables real-time support for process workers during the manufacturing process. Using the tool, a worker can recognize the factors that can cause quality problems in the process in advanc, Tiivistelmä Teollisuus on siirtynyt yhä enenemissä määrin kohti dataohjautuvaa liiketoimintaa. Dataa on tarjolla niin paljon, että ongelmana ei ole sen saatavuus vaan se mitä siitä saadaan irti. Datasta louhittu tieto auttaa kontrolloimaan teollisuusprosesseja tehokkaammiksi ja automaattisemmiksi sekä parantamaan tuotteen laatua löytämällä syyt huonon laadun takaa ja näin parantamaan koko tehtaan tuottoa ja kilpailukykyä. Tämä väitöskirja auttaa ymmärtämään tilastollisten koneoppimismenetelmien tärkeyden teollisuusprosesseja kehitettäessä. Näiden menetelmien avulla pystymme ennakoimaan tulevaisuutta historian avulla ja kasvattamaan dataohjautuvan päätöksenteon tukea prosesseissa. Tämän väitöskirjan sovellusalueena on terästeollisuus. Työ antaa työkaluja siihen miten AI sovellus voidaan toteuttaa onnistuneesti teollisuudessa askel askeleelta. Lisäksi työssä esitellään menetelmiä joiden avulla voidaan löytää juurisyitä huonolle laadulle ja näin kehittää prosessia tuottamaan laadukkaampia tuotteita. Koska teollisuuden haastavista olosuhteista kerätty data ei koskaan ole täydellistä, on tässä työssä kehitetty myös laatumalli, joka pystyy hyödyntämään puuttuvia tietoja. Tässä työssä käytetään koneoppimismenetelmiä, joiden tarkoituksena on prosessoida dataa ja kehittää datamukautuvia laatumalleja, joiden avulla voidaan tehdä ennustuksia haluttavasta laatuominaisuudesta, ja näin tukea prosessissa työskentelevien ihmisten päätöksentekoa. Työssä on käytetty sekä ohjattuja-, että osittain-ohjattuja koneoppimismenetelmiä sekä mallien läpinäkyyvyyttä lisääviä selittäviä koneoppimismenetelmiä. Näiden menetelmien avulla on kehitetty ennustavia laatumalleja neljän erilaisen teräsprosessia kuvaavaan datasetin avulla. Kehitetyt laatumallit on implementoitu teollisuutta varten kehitettyyn älykkään päätöksenteon työkaluun, joka mahdollistaa reaaliaikaisen tuen teollisuusprosessissa työskenteleville henkilöille. Työkalun avulla työntekijä voi jo etukäteen tunnistaa tekijät, jotka voiva

Sensor devices are becoming omnipresent, supplying data to a wide range of applications. In the building sector, sensors along with other information sources provide the basis for smart building functionalities. Predicting energy loads and inferring occupancy status of spaces are important tasks that promote energy efficiency and user comfort in buildings. For them, as for many other smart building applications, machine learning modelling utilizing sensor data is commonly applied. This article builds understanding of the environment where this kind of machine learning models have to operate by bringing up properties and quality aspects of the public building data provided by indoor sensor devices. This is done by performing a thorough case study on two real life data sets from university campus buildings located in different climates and applying very different sensor network settings. Outcomes include information about heterogeneity, correlations and temporal patterns present in sensor data, and show the need of the building field for better acknowledging the quality deficiencies that sensor data have. Our results aid in assessing and improving the quality of sensor-based indoor data utilized in machine learning modelling, in evaluating whether a data set is representative enough to build a model that is robust under changing conditions in the building, and in choosing an appropriate number of sensors per space when building an indoor wireless sensor network.

This article demonstrates the use of data mining methods for evidence-based smart decision support in quality control. The data were collected in a measurement campaign which provided a new and potential quality measurement approach for manufacturing process planning and control. In this study, the machine learning prediction models and Explainable AI methods (XAI) serve as a base for the decision support system for smart manufacturing. The discovered information about the root causes behind the predicted failure can be used to improve the quality, and it also enables the definition of suitable security boundaries for better settings of the production parameters. The user’s need defines the given type of information. The developed method is applied to the monitoring of the surface roughness of the stainless steel strip, but the framework is not application dependent. The modeling analysis reveals that the parameters of the annealing and pickling line (RAP) have the best potential for real-time roughness improvement.

Non-metallic inclusions, especially large or clustered inclusions, in steel are usually harmful. Thus, the microscopic analysis of test specimens is an important part of the quality control. This steel purity analysis produces a large amount of individual inclusion information for each test specimen. The interpretation of the results is laborious and the comparison of larger product groups practically impossible. The purpose of this study was to develop an easy-to-use tool for automatic interpretation of the SEM analysis to differentiate clustered and large inclusions information from the manifold individual inclusion information. Because of the large variety of the potential users, the tool needs to be applicable for any steel grade and application, both for liquid and final product specimen, to analyse automatically steel specimen inclusions, especially inclusion clusters, based on the INCA Feature program produced data from SEM/EDS. The developed tool can be used to improve the controlling of the steel purity or for automatic production of new inclusion cluster features that can be utilised further in quality prediction models, for example.

Advances in technology and data analysis provide rich opportunities for developing intelligent environments assisting their inhabitants, so-called smart environments or smart spaces. Enhanced with technology, sensors, user interfaces, and various applications, such smart spaces are capable of recognizing users and situations they are in, react accordingly, e.g., by providing certain services or changes to the environment itself. Therefore, smart space solutions are gradually coming to different application domains, each with corresponding specific characteristics. In this article, we discuss our experiences and explore the challenges of a long-term real-world Internet of Things (IoT) deployment at a University campus. We demonstrate the technical implementation and data quality issues. We conduct several studies, from data analysis to interaction with space, utilizing the developed infrastructure, and we also share our actions to open the data for education purposes and discuss their outcomes. With this article, we aim to share our experience and provide real-world lessons learned when building an open, multipurpose, publicly used smart space at a University campus.

Advances in technology equip traffic domain with instruments to gather and analyse data for safe and fuel-efficient traveling. In this article, we elaborate on the effects that taxi drivers’ route selection has on fuel efficiency. For this purpose, we fuse real driving behaviour data from taxi cabs, weather, digital map, and traffic situation information to gain understanding of how the routes are selected and what are the effects in terms of fuel-efficiency. Analysis of actually driven trips and their quickest and shortest counterparts is conducted to find out the fuel-efficiency consequences on route selection. The judgments are used for developing a fuel-consumption model, exploring further the route characteristics and external factors affecting fuel consumption.

This article presents a statistical prediction model-based intelligent decision support tool for center line deviation monitoring. Data mining methods enable the data driven manufacturing. They also help to understand the manufacturing process and to test different hypotheses. In this study, the original assumption was that the shape of the strip during the hot rolling has a strong effect on the behaviour of the steel strip in Rolling, Annealing and Pickling line (RAP). Our goal is to provide information that enables to react well in advance to strips with challenging shape. In this article, we show that the most critical shape errors arising in hot rolling process will be transferred to critical errors in RAP-line process as well. In addition, our results reveal that the most critical feature characterizes the deviation better than the currently used criterion for rework. The developed model enables the user to understand better the quality of the products, how the process works, and how the quality model predicts and performs.

The purpose of this study was to develop an innovative online supervisor system to assist the operators of an industrial manufacturing process in discovering new solutions for improving both the products and the manufacturing process itself. In this paper, we discuss the requirements and practical aspects of building such a system and demonstrate its use and functioning with different types of statistical modelling methods applied for quality monitoring in industrial applications. The two case studies presenting the development work were selected from the steel industry. One case study predicting the profile of a stainless steel strip tested the usability of the tool offline, while the other study predicting the risk of roughness of a steel strip had an online test period. User experiences from a test use period were collected with a system usability scale questionnaire.

Physical exercise can improve sleep quality. However, how to perform physical exercise to achieve the best possible improvements is not clear. In this article, we build predictive models based on volume real data collected from wearable devices to predict the sleep efficiency related to users’ daily exercise information. As far as we know, this is the first study to investigate insights of prediction of sleep efficiency from volume physical exercise data collected from real world.

Heart rate is a good measure for physical exercise as it accurately reflects exercise intensity and is easy to measure. If the heart rate response to a complete exercise session is predicted beforehand, information related to the exercise can be inferred, such as exercise intensity and calorie consumption. While most current heart rate prediction models are developed and tested for the scenarios of indoor running exercise or low running speed exercise, we adopt a nonlinear Ordinary Differential Equation (ODE) model for complete outdoor running exercise sessions to predict the heart rate response and identify the parameters of the model with machine learning algorithms. The proposed model enables us to predict a complete outdoor running exercise session instead of predicting the heart rate for a short duration. Model validation is carried out both on the training and testing sets. Our results show that the proposed model captures very stable prediction performance.

Data are crucial to support decision making. If data have low veracity, decisions are not likely to be sound. Internet of Things (IoT) generates big data with inaccuracy, inconsistency, incompleteness, deception, and model approximation. Enhancing data veracity is important to address these challenges. In this article, we summarize the key characteristics and challenges of IoT, which influence data processing and decision making. We review the landscape of measuring and enhancing data veracity and mining uncertain data streams. Moreover, we propose five recommendations for future development of veracious big IoT data analytics that are related to the heterogeneous and distributed nature of IoT data, autonomous decision-making, context-aware and domain-optimized methodologies, data cleaning and processing techniques for IoT edge devices, and privacy preserving, personalized, and secure data management.

This paper presents an AI based method for improving poorly performing quality prediction models. The method improves automatically the usability of the low quality alarm predictions in the web based quality monitoring tool that provides decision support for users. The tool enables the utilization of the models that suffer from the lack of information because of a long time gap to the predicted future. The reliability of the presented alarms in a monitoring tool will be improved by reducing the amount of false alarms.

Quality control is an essential part of manufacturing, and the different properties of the products can be tested with standardized methods. If the decision of qualification is based on only one test specimen representing a batch of products, the testing procedure is quite straightforward. However, when the measured property has a high variability within the product, as usual, several test specimens are needed for the quality verification. When a quality property is predicted, the response value of the model that most effectively finds the critical observations should naturally be selected. In this thesis, it has been shown that LIB-transformation (Larger Is Better) is a suitable method for multiple test samples, because it effectively recognizes especially the situations where one of the measurements is very low. The main contribution of this thesis is to show how to model quality of phenomena that consist of several measurement samples for each observation. The process contains several steps, beginning from the selection of the model type. Prediction of the exceedance probability provides more information for the decision making than that of the mean. Especially with the selected application, where the quality property has no optimal value, but the interest is in adequately high value, this approach is more natural. With industrial applications, the assumption of constant variance should be analysed critically. In this thesis, it is shown that exceedance probability modelling can benefit from the use of an additional variance model together with a mean model in prediction. The distribution shape modelling improves the model further, when the response variable may not be Gaussian. As the proposed methods are fundamentally different, the model selection criteria have to be chosen with caution. Different methods for model selection were considered and commented, and EPS (Exceedance Probability Score) was chosen, because it is most suitable for probability pr, Tiivistelmä Laadunvalvonnalla on keskeinen rooli teollisessa tuotannossa. Valmistettavan tuotteen erilaisia ominaisuuksia mitataan standardin mukaisilla testausmenetelmillä. Testi on yksinkertainen, jos tuotteen laatu varmistetaan vain yhdellä testikappaleella. Kun testattava ominaisuus voi saada hyvin vaihtelevia tuloksia samastakin tuotteesta, tarvitaan useita testikappaleita laadun varmistamiseen. Tuotteen laatuominaisuuksia ennustettaessa valitaan malliin vastemuuttuja, joka tehokkaimmin tunnistaa laadun kannalta kriittiset havainnot. Tässä väitöskirjassa osoitetaan, että LIB-transformaatio (Large Is Better) tunnistaa tehokkaasti erityisesti tilanteet, joissa yksi mittauksista on hyvin matala. Tämän väitöskirja vastaa kysymykseen, kuinka mallintaa laatua, kun tutkittavasta tuotteesta tarvitaan useita testinäytteitä. Mallinnusprosessi koostuu useista vaiheista alkaen mallityypin valinnasta. Alitusriskin mallinnuksen avulla saadaan enemmän informaatiota päätöksenteon tueksi perinteisen odotusarvomallinnuksen sijaan, etenkin jos laatutekijältä vaaditaan vain riittävän hyvää tasoa optimiarvon sijaan. Teollisissa sovelluksissa ei voida useinkaan olettaa, että vasteen hajonta olisi vakio läpi prosessin. Tässä väitöskirjassa osoitetaan että alitusriskin ennustamistarkkuus paranee, kun odotusarvon lisäksi mallinnetaan myös hajontaa. Jakaumamuodon mallilla voidaan parantaa ennustetarkkuutta silloin, kun vastemuuttuja ei noudata Gaussin jakaumaa. Koska ehdotetut mallit ovat perustaltaan erilaisia, täytyy myös mallin valintakriteeri valita huolella. Työssä osoitetaan, että EPS (Exceedance Probability Score) toimii parhaiten käytetyillä todennäköisyyttä ennustavilla malleilla. Tässä väitöskirjassa osoitetaan, että erityisesti silloin kun tuotantoprosessi on monimuotoinen ja laatumuuttujan jakaumamuoto on haastava, mallinnuttaminen hyötyy hajontamallin käytöstä, ja tuloksia voidaan parantaa jakaumamuodon mallilla.

During the last few years, there have been debates over what is context and how computers should act upon it. Two disparate camps of thought can be recognized. First, Realism, having its roots in natural sciences, believes that contexts exist out there and that, if properly instrumented and programmed, computers can correctly recognize and adapt to them. Second, Constructivism, having its roots in human and social sciences, believes that contexts are human creations, mental and social, and that computers ought to provide resources for managing them. We reveal some fundamental differences between the two in three different application domains. We show that despite the deep-going controversies, both camps benefit from considering the alternative approach and a middle ground can be found., QC 20141128