IEEE Transactions on Knowledge and Data Engineering

— The Unified Modeling Language (UML) supports the capture and representation of conceptual semantics in analysis and modeling of arbitrary domains of discourse. It is now being used for modeling of database and agent-oriented applications in addition to its use with object-oriented software. An important activity in conceptual modeling is the capture and expression of domain constraints on the underlying data and object states. UML supports the capture of declarative constraints both directly in the graphical class diagram notation and in the separately denoted Object Constraint Language (OCL) syntax. In this paper, we present a set of constraint patterns that have been formalized for capturing a broader category of semantic constraints directly using the class diagram notation. The objective of this work is to enrich UML static models for greater precision in conceptual modeling applications by formalizing the syntax and semantics of these constraint patterns, thus giving analysts a tool for capturing these common patterns directly in the UML class diagram with a minimum of additional notation—without requiring a separate expression of constraints in the adjunct OCL syntax. We present a taxonomy for propositional constraints, discuss their semantics through the use of logic truth tables and Karnaugh maps, and illustrate their usage in UML class diagrams by way of an example from the Unix security domain.

Citations in the corpus

• Learning to Run Heuristics in Tree Search (2017, methodology): “Data points with label y H = 1 are heavily weighted in the LR loss function to account for the extreme class imbalance we encounter [He and Garcia, 2009], as can be seen in column “Success Rate” of Table 3.”
• A learning theory approach to non-interactive database privacy (2008, background): “[XWG10] gave an algorithm for releasing range queries, which extends the class of constantdimensional interval queries which we consider in this paper.”
• Evolutionary stratified training set selection for extracting classification rules with trade off precision-interpretability (2007, methodology): “When the decision tree is going to be applied in domains where its character predictive and descriptive is important, the simplicity of the decision tree is a key factor [2].”
• Incremental Learning in Deep Convolutional Neural Networks Using Partial Network Sharing (2017, methodology): “But previous works [17] used this method to learn a different dataset using the frozen parameters as fixed feature extractors.”
• Data clustering: a review (1999, background): “Several journals have had recent special issues on data mining [Cohen 1996, Cross 1996, Wah 1996].”
• Anomaly Detection in Streams with Extreme Value Theory (2017, unknown): “The main approaches are distance based [7], nearestneighbor based [11] or clustering based [14] and are very well detailed in [13].”
• Discretization Techniques: A recent survey (2006, methodology): “The goal of the CAIM algorithm [26] is to maximize the class-attribute interdependence and to generate a (possibly) minimal number of discrete intervals.”
• Inherent Trade-Offs in the Fair Determination of Risk Scores (2016, background): “Kamiran and Calders [21] and Hajian and Domingo-Ferrer [18] seek to modify datasets to remove any information that might permit discrimination.”
• Unsupervised Cross-Modality Domain Adaptation of ConvNets for Biomedical Image Segmentations with Adversarial Loss (2018, background): “, fine-tuning the models learned on source domain with extra labeled data from the target domain [Pan and Yang, 2010].”
• DynGEM: Deep Embedding Method for Dynamic Graphs (2018, methodology): “Note that unlike [ Zhu et al. , 2016 ] we do not impose explicit regularizers to keep the embeddings at time steps t − 1 and t close.”
• Non-negative Matrix Factorization on Manifold (2008, methodology): “Two metrics, the accuracy ( AC) and the normalized mutual information metric ( MI) are used to measure the clustering performance [29][5].”