]> KIDS Knowledge Intensive Database System (KIDS) Knowledge Intensive Database System (KIDS) is a formal framework for decision making based on the OODA loop that provides for (1) Categorization of data and reasoning processes; (2) State management of data, reasoning and process; (3) Process model for situation awareness and decision making Knowledge Intensive Database System (KIDS) is a formal framework for decision making based on the OODA loop that provides for (1) Categorization of data and reasoning processes; (2) State management of data, reasoning and process; (3) Process model for situation awareness and decision making International Association for Ontology and Application 21 July 2016 https://ken.baclawski.com/kids.owl Version 1.0 Kenneth Baclawski (ken@baclawski.com), Northeastern University Kenneth Baclawski (ken@baclawski.com), Northeastern University Eric S. Chan (eric.s.chan@oracle.com), Oracle Dieter Gawlick (dieter.gawlick@oracle.com), Oracle Adel Ghoneimy (adel.ghoneimy@oracle.com), Oracle Kenny Gross (kenny.gross@oracle.com), Oracle Zhen Hua Li (zhen.liu@oracle.com), Oracle Xing Zhan (zhang.xing1@husky.neu.edu), Northeastern University Copyright © Kenneth Baclawski 2016. All Rights Reserved. A collection of activities for decision making. Normally, the collection of activities consists of a sequence of loop iterations where each iteration is in the order classify, assess, resolve and enact, but it is possible to combine the assess and resolve activities in a single step. It is also possible for the activities to branch and merge, forming a directed graph of activities. A process that uses the information in a situation to produce another situation. Logical reasoning is expressed in terms of rules. Many techniques are available for probabilistic reasoning. A reasoning process that extracts data relevant to a goal or purpose. Modern sensor technology produces very large amounts of data, only a small part of which will be relevant to a particular goal. The activity is called filtering. Relevance reasoning includes both extraction from raw data as the initial step in a decision making process and detection of relevant new or modified information later in the process. A reasoning process that uses data acquired externally by observations, such as by using sensors, or internally from previous decision cycles. The situation produced by a classification process is a perception. Classification is usually implemented by deductive reasoning, but inductive reasoning is also common. The wide variety of classification techniques includes Support Vector Machines, Naïve Bayesian Networks, Neural Networks, Clustering, Association Rules, Decision Trees, Multivariate State Estimation Technique, Cognitive computing, etc. A reasoning process that uses perceptions and produces hypotheses relevant to the decision making goal. The hypotheses are also called diagnoses. Assessment is typically implemented by abductive reasoning. Examples of techniques include Bayesian Belief Networks, Expectation Maximization, Least-Squares Optimization and Regression of solutions for inverse problems. A reasoning process that uses hypotheses and produces directives that represent the decision that was made to address the goal. Resolution involves making decisions under the uncertainty of outcomes by considering the relative merit of the different outcomes and the associated payoffs/costs. Resolution techniques include Bayesian Belief Network extended with decision and payoff/cost nodes, known as Influence Diagrams, Dempster-Shafer theory, Decision Trees, and Prognosis of Remaining Useful Life. A directive typically includes a plan of action in the form of a workflow or process. The plan may achieve the goal. If the plan does not fully achieve the goal, then the plan will either involve additional observations or give a reason why the goal should not be achieved. A reasoning process that combines the assessment and resolution steps. It differs from performing assessment and resolution separately in not generating hypotheses. For example, in the healthcare domain, a practitioner will often treat a patient without determining the actual disease. A reasoning process that performs the directives, thereby affecting the world in some way. Enactment involves control structures of various kinds, usually encoded in scripts, plans, schedules, BPEL workflows, and business processes in BPMN. Enactment often includes a request for additional observations. Information relevant to achieving a goal. The kinds of data are classified by when they are used in the decision loop. The information relevant to a goal, acquired externally by observations, such as by using sensors, or internally from previous decision cycles. Compact representations of facts. Perceptions are generally at a higher level, than raw facts. For example, a fact might be an MRI scan of a patient, while a perception based on the MRI scan would identify objects in the MRI scan that are physiologically relevant to the goal of diagnosing the patient. The potential causes that explain perceptions. Hypotheses are also called diagnoses. A plan for achieving the goal either fully or partially, or a reason why the goal is not desirable. A plan is usually encoded in scripts, plans, schedules, BPEL workflows, and business processes in BPMN. Enactment often includes a request for additional observations. A period of time. The period will always have a starting time. The ending time can be an actual time or it will indicate some reason why the ending time is not specified, such as that the period of time proceeds indefinitely into the future, or that the period of time is instantaneous so the ending time is not applicable. A specification of the requirements for data used by an activity or the characteristics of data produced by an activity. A flexible relation notion from situation theory that includes, flexible schema data, categorical features, as well as other data and object properties. It is not limited to tables from the relational model. The execution context of a relevance reasoning process. This class is defined as the activities that use relevance reasoning processs. The execution context of a classification process. This class is defined as the activities that use classification processs. The execution context of an assessment process. This class is defined as the activities that use assessment processs. The execution context of a resolution process. This class is defined as the activities that use resolution processs. The execution context of a process that combines assessment and resolution. This class is defined as the activities that use processs that combine assessment and resolution. The execution context of an enactment process. This class is defined as the activities that use enactment processs. A situation that supports relevance reasoning activities. A situation that supports classification activities. A situation that supports assessment activities. A situation that supports activities that combine assessment and resolution. A situation that supports resolution activities. A situation that supports enactment activities. A situation that supports facts. A situation that supports perceptions. A situation that supports hypotheses. A situation that supports directives. The situation used by an activity to produce another situation. The situation produced by an activity using another situation. The infons that an activity is using. The infons that an activity is producing. The reasoning process used by an activity. The CARE-Loop containing a directed graph of situations. The collaborative workspace for a CARE-Loop. The time interval during which data are valid. The next focal situation in a valid time series. The next focal situation in a transaction time series. The guard that specifies the input requirements of an activity. The guard that specifies the characteristics of the output of an activity. The starting time of a time interval. The ending time of a time interval. It can be an actual time or it will indicate some reason why the ending time is not specified, such as that the period of time proceeds indefinitely into the future, or that the period of time is instantaneous so the ending time is not applicable.