Umeå University's logo

The focus of this paper is the semi-quantitative approach to risk analysis in HACCP (Hazard Analysis and Critical Control Points), where we show how alternative computations in the risk matrix, combining probability and impact, can be provided based on algebraic and many-valued logical techniques. Doing so, we further show how applying such computations requires being formal concerning underlying information structures, which in turn enables being formal concerning functional representation of mappings between information structures appearing within risk analysis in risk management. Our mathematical algebraic framework also enables a more formal treatment of the duality between threat and opportunity.

In this paper we argue in favor of an intertwining of information and processes within enterprise information management, with the objective to show how an information model is a torso without data being localizable within a process, and similarly, a process model is a torso if it contains neither data or data placeholders as they connect across subprocesses within the overall process. The paper contributes with an example scenario is drawn from service management within the mining industry. We give an overview of enterprise information views and we show how information is localizable within process views based on BPMN (Business Process Modeling Notation) and its Pools and Lanes. The outcome is design support for development of digitization and digitalization approaches including analytics and artificial intelligence to monitor production processes and prescribe optimal operation to optimize efficiency, avoid unplanned stops, and decrease environmental impacts. Given the complexities of mining operations, we show how further development is required for standards for process and information models.

In the set QX of Q-fuzzy sets over a set X, members appear in X and memberships in Q. In the set (QbQe)X of QbQe-fuzzy sets over X, a membership of X is in QbQe and a membership of Qe is in Qb. Thus, when moving to deal with fuzzy sets of higher types, memberships become members having memberships. The roles of members and memberships may become confusing, particularly in type reductions, unless we clearly separate members from memberships, where ever they appear. In order to avoid confusion, and to enable a recursive approach to dealing with fuzzy sets of higher types and their reductions, the paper presents a functional approach to reductions of QbQe-fuzzy sets, where Qb and Qe are quantales. Using constructions of type n, n≥2, fuzzy sets with memberships given from quantales, we study three different approaches to type reducers which are based on choice functions. A brief discussion on quantale-valued geometric mean as a basis of type reducers is also given.

In this paper we introduce a systemic approach for predictive and prescriptive maintenance framed within the larger system of systems, as exemplified by a use case in mining. Developments are presented as systematic, while there is a focus on improved availability modeling using the time usage model with a corresponding UML/SysML StateMachine representation. Data becomes connected with a more elaborated definition of time, with failure modes and analytics in reliability engineering being supported by improved underlying information structures.

In this paper, we open up more possibilities to define higher types of fuzzy sets in a mixed way. In doing so, we show that there are essentially two alternative definitions for fuzzy sets of higher types, one of which is widely adopted in the literature, and one where the unit interval, or subsets thereof, is used to represent membership values. The other alternative definition opens up new perspectives for the use of fuzzy sets of higher types, and it promotes the use of other algebraic structures of sets of membership values, where quantales are seen as particularly useful, also in applications. The paper also underlines the importance of making distinctions between “computing with fuzzy” and “fuzzy computing” and understanding the difference between “logic with fuzzy” and “fuzzy logic”.

In this chapter the authors compare assessment techniques in education and health. These examples are drawn from assessment for learning in secondary school as compared with assessment of functioning in care of older people. Content in these respective assessment environments is obviously quite different, but there is some overlap, in particular related to social circumstances. The authors further argue that the assessment structure and formal analysis techniques are and even should be quite similar. These two domains of assessments share not only a common information and terminological structure, but also a common assessment framework (CAF) which is shown to be extendable by process detail. In this focus on assessment in health, particularly in active and healthy ageing, the model is intended to support the integration of information and processes (decision and intervention pathways) supporting various types of providers.

This paper analyzes data from coronary artery bypass grafting (CABG) using decision functions to represent rules. The data was collected at the University Hospital in Umeå, Sweden. The data contains pre-, intra-, and postoperative detail from 2975 heart operations during 1993-96. Each instance is represented by 14 preoperative variables, 4 intraoperative variables, and 9 postoperative variables. A logical method is used to predict the postoperative variables using preoperative variables. First, each postoperative variable is represented as a decision functions of preoperative variables. Then, for each postoperative variable, a minimal set of preoperative variables is derived. And finally, each postoperative variable is represented by a minimum set of rules using preoperative variables. With this method we can predict postoperative outcome, where prediction using preoperative data only is of particular interest e.g. for surgery scheduling.

Introduction: Data from mHealth apps can provide valuable information on rhinitis control and treatment patterns. However, in MASK-air®, these data have only been analyzed cross-sectionally, without considering the changes of symptoms over time. We analyzed data from MASK-air® longitudinally, clustering weeks according to reported rhinitis symptoms. Methods: We analyzed MASK-air® data, assessing the weeks for which patients had answered a rhinitis daily questionnaire on all 7 days. We firstly used k-means clustering algorithms for longitudinal data to define clusters of weeks according to the trajectories of reported daily rhinitis symptoms. Clustering was applied separately for weeks when medication was reported or not. We compared obtained clusters on symptoms and rhinitis medication patterns. We then used the latent class mixture model to assess the robustness of results. Results: We analyzed 113,239 days (16,177 complete weeks) from 2590 patients (mean age ± SD = 39.1 ± 13.7 years). The first clustering algorithm identified ten clusters among weeks with medication use: seven with low variability in rhinitis control during the week and three with highly-variable control. Clusters with poorly-controlled rhinitis displayed a higher frequency of rhinitis co-medication, a more frequent change of medication schemes and more pronounced seasonal patterns. Six clusters were identified in weeks when no rhinitis medication was used, displaying similar control patterns. The second clustering method provided similar results. Moreover, patients displayed consistent levels of rhinitis control, reporting several weeks with similar levels of control. Conclusions: We identified 16 patterns of weekly rhinitis control. Co-medication and medication change schemes were common in uncontrolled weeks, reinforcing the hypothesis that patients treat themselves according to their symptoms.

This chapter shows how assessment values often are viewed as part of a numeric scale, where in fact the fact values are understood more as symbolic in an ordinary scale. Computing with numeric and symbolic values is different, and challenging in particular if and when computations have to be intertwined. Computing with numeric values involves traditional applied and numerical mathematics, whereas computing with symbolic values involves logic and algebra. In this chapter the authors illuminate these aspects using examples from classification of functioning. 

In this chapter, the authors describe an architecture for reinforcement learning in social media marketing. The rule bases used for action selection within the architecture build upon many-valued (fuzzy) logic. Action evaluation and internal learning is based on neural network like structures. In using variables measuring the effect of advertising, we must understand direction of influence between advertiser, owning the content of the advertisement, and advertisee, as the target of an advertisement, and as facilitated by social media marketing. Examples are drawn from Facebook marketing.