Behavior refers to the action, reaction or property of an entity, human or otherwise, to situations or stimuli in its environment. It is ubiquitous in real life and has been highlighted as a key component in business intelligence and complex problem-solving. The in-depth analysis of human behavior has been increasingly recognized as a crucial means for understanding and disclosing interior driving forces and intrinsical cause-effects on business and social applications in handling many challenging issues, including behavior modeling and analysis in virtual organizations, web community analysis, counter-terrorism, fraud detection, and customer relationship management. To the best of our knowledge, behavior modeling and analysis have been extensively studied by researchers in different disciplines, e.g. Psychology, Economics, Mathematics, Engineering, and Information Science. From those diverse perspectives, there are widespread and long-standing explorations on behavior studies , such as outlier mining of trading behaviors, telecom churn analysis, periodic behavior analysis, social network analysis, sequence analysis, multivariate time series, and interactive process modeling, etc. However, all the above emerging methods suffer from the following common issues and problems to different extents: (1) Existing behavior modeling approaches have too many styles and forms according to distinct situations, which is troublesome for cross-discipline researchers to follow. (2) Traditional behavior analysis relies on implicit behavior and explicit business appearance, often leading to ineffective and limited understanding on business and social activities. (3) Complex coupling relationships among behaviors are often ignored or only weakly addressed, which fails to provide a complete understanding of the underlying problems and their comprehensive solutions. (4) Current research usually overlooks the checking of behavior interaction modeling, which weakens the soundness and robustness of models built for complex behavior applications. (5) Most of the current mining and learning algorithms follow the classic assumption of independence and identical distribution (i.e. IIDness), but this is too strong to match the reality and complexities in real-world applications.

With the deepening and widening of social/business intelligences and their networking, the concept of behavior is in great demand to be consolidated and formalized to deeply scrutinize native behavior intention, lifecycle and impact on complex problems and business issues. In the real-world applications, group behavior interactions, such as multi-robot teamwork and group communications in social networks, are widely seen in natural, social and artificial behavior-related problems. The verification of behavior modeling is further desired to assure the reliability and stability. In addition, complex behavioral and social applications often exhibit strong explicit or implicit coupling relationships and heterogeneity both between their entities and properties. They can not be abstracted or weakened to the extent of satisfying the IIDness assumption. These characteristics greatly challenge the current behavior-related analysis approaches. Moreover, it is also very difficult to model, analyze and check behaviors coupled with one another due to the complexity from data, domain, context and impact perspectives. Based on the above research limitations and challenges, this thesis reports state-of-the-art advances and our research innovations in modeling, analyzing and learning coupled behaviors, which constitute the coupled behavior informatics. Coupled behaviors refer to the activities of one to many actors who are associated with each other in terms of certain relationships. They are categorized as qualitative coupled behaviors and quantitative coupled behaviors, depending on whether the behavior involved is qualified by actions or quantified by properties. Coupled behavior modeling is to develop representation and modeling mechanisms to capture behavior characteristics, intrinsic and contextual properties of behaviors, as well as the coupling relationships therein. Coupled behavior analysis is to propose effective techniques for emergent domains in analyzing coupled behaviors and their properties. Coupled behavior learning is to identify clusters and patterns among the quantitative behavior entities and networks.

In terms of the qualitative coupled behavior modeling and analysis, we propose an Ontology based Behavior Modeling and Checking (OntoB for short) system to explicitly represent and verify complex behavior relationships, aggregations, and constraints. The OntoB system provides both a visual behavior ontology and an abstract behavior tuple to capture behavioral elements, as well as building blocks to formalize various intra-coupled interactions (behaviors conducted by the same actor) via transition systems, and inter-coupled behavior aggregations (behaviors conducted by different actors) from temporal, inferential and party-based perspectives. OntoB converts a behavior-oriented application into a transition system and temporal logic formulae for further verification and refinement. We demonstrate the effectiveness of the OntoB system in modeling multi-robot behaviors and their interactions in the Robocup soccer competition game. We show the OntoB system can effectively model complex behavior interactions, verify and refine the modeling of complex group behavior interactions in a sound manner. With regard to the quantitative coupled behavior analysis and learning, we carry out explorations on three tasks: the numerical coupled behavior analysis, the categorical coupled behavior analysis, the coupled behavior ensemble learning. All the proposed methods and algorithms below are under the non-IIDness assumption of entities or properties or both of them, which caters for the intrinsical essence of real-world problems and applications. In the following, object and the entity of coupled behaviors are interchangeable, attributes and base clusterings indicate the properties of coupled behaviors. For the first part, we introduce a framework of the coupled property analysis to capture the global dependency of continuous attributes. Such global couplings integrate the intra-coupled interaction within an attribute (i.e. the correlations between attributes and their own powers) and inter-coupled interaction among different attributes (i.e. the correlations between attributes and the powers of others) to form a coupled representation for numerical entities by the Taylor-like expansion. This work makes one step forward towards explicitly addressing the global interactions of continuous attributes, verified by the applications in data structure analysis, clustering and classification. Substantial experiments on 13 UCI data sets demonstrate the coupled representation can effectively capture the global couplings of attributes and outperforms the traditional way, supported by statistical analysis. For the second part, we present an efficient data-driven similarity learning approach that generates a coupled attribute similarity measure for nominal entities with property couplings to capture a global picture of attribute similarity. It involves the frequency-based intra-coupled similarity within a property and the inter-coupled similarity upon value co-occurrences between properties, as well as their integration on the entity level. In particular, four measures are designed for the inter-coupled similarity to calculate the similarity between two categorical values by considering their relationships with other attributes in terms of power set, universal set, join set and intersection set. The theoretical analysis reveals the equivalent accuracy and superior efficiency of the measure based on the intersection set, particularly for large-scale data sets. Substantial experiments on 20 UCI data sets verify the theoretical conclusions. In addition, intensive experiments of data structure, clustering and classification algorithms incorporating the coupled dissimilarity metric achieve a significant performance improvement on stateof- the-art measures and algorithms on 12 UCI data sets and bibliographic data, which is confirmed by the statistical analysis. The experiment results show that the proposed coupled attribute similarity is generic, and can effectively and efficiently capture the intrinsical and global interactions within and between properties for especially large-scale categorical data sets. For the last part, we explicate the non-IIDness between base clusterings and between objects in clustering ensembles, and put forward a framework for coupled clustering ensembles (CCE). CCE not only considers but also integrates the coupling relationships between properties and between entities. Specifically, we examine both the intra-coupling within one base clustering (i.e. cluster label frequency distribution) and the inter-coupling between different base clusterings (i.e. cluster label co-occurrence dependency). Furthermore, we engage both the intra-coupling between two objects by aggregating the interactions of base clusterings and the inter-coupling among other objects by exploring their neighborhood domains. This is the first work which explicitly addresses the non-IIDness issue in clustering ensembles, verified by the application of such couplings in three types of consensus functions: clustering-based, object-based and cluster-based. Substantial experiments on two synthetic and nine UCI data sets demonstrate that the CCE framework can effectively capture the interactions embedded in base clusterings and objects with higher clustering accuracy, stability, and robustness compared to eleven state-of-the-art techniques, supported by statistical analysis. Additionally, we verify that the final clustering quality is dependent on the data characteristics (i.e. the quality and consistency) of base clusterings.

Finally, we provide a consolidated understanding of coupled behaviors by summarizing the qualitative and quantitative aspects, extract the multilevel couplings embedded in coupled behaviors, and then formalize a coupled behavior algebra at its preliminary stage. Several open research issues and opportunities for this algebra, such as the definition of computing operators and the construction of a coupled behavior space, are pointed out. Under varying backgrounds and scenarios, our proposed methods and algorithms for coupled behavior informatics are evidenced to outperform stateof- the-art approaches via theoretical analysis or empirical studies or both of them. All these outcomes have been accepted by top conferences, and some follow-up work has also been recognized by research peers. Thus, the coupled behavior informatics is a promising though wholly new research topic with a lot of attractive opportunities for further exploration and development.