Advances in Conceptual Modeling

Sharing economy is prevailing. The network of cars and shared bicycles is convenient for people to travel. We investigate the issue of value-added service based on crowdsourcing for campus shuttles. We can provide diverse services between users by solving matching problems. The service concludes positioning and location services, requesting designating. The efficient incentive mechanisms make the shuttle bus transportation parcel convenient. We use KNN algorithm to establish KD tree to index different parcels nodes. In our app demo, we show how the application execute and how to improve the user experience who involve the orders.

Entity-oriented information access is becoming a key enabler for next-generation information retrieval and exploration systems. Previously, researchers have demonstrated that knowledge graphs allow the exploitation of semantic correlation among entities to improve information access. However, less attention is devoted to user interfaces of tablets for exploring knowledge graphs effectively and efficiently. In this paper, we design and implement a system called SEED to support entity-oriented exploratory search in knowledge graphs on tablets. It utilizes a dataset of hundreds of thousands of film-related entities extracted from DBpedia V3.9, and applies the knowledge embedding derived from a graph embedding model to rank entities and their relevant aspects, as well as explaining the correlation among entities via their links. Moreover, it supports touch-based interactions for formulating queries rapidly.

The popularity of smartphones and the advances in location-acquisition technologies witness the development in the research of human mobility. This demo shows a relationship-discovery system of mobile users from their spatio-temporal trajectories. The system first matches all the access device IDs to places of interest (POI) on the map, and then finds out the access device IDs visited by more than one phone frequently or regularly. For these users, a model of historical spatio-temporal trajectories analysis combined with web browsing behavior is proposed to discover the relationship among them. A large-scale real-life mobile data set has been used in constructing the system, the performance of which is evaluated to be effective, efficient and user-friendly.

Entity type completion in Knowledge Bases (KBs) is an important and challenging problem. In our recent work, we have proposed a hybrid framework which combines the human intelligence of crowdsourcing with automatic algorithms to address the problem. In this demo, we have implemented the framework in a crowdsourcing-based system, named Crowd-Type, for fine-grained type completion in KBs. In particular, Crowd-Type firstly employs automatic algorithms to select the most representative entities and assigns them to human workers, who will verify the types for assigned entities. Then, the system infers and determines the correct types for all entities utilizing both the results of crowdsourcing and machine-based algorithms. Our system gives a vivid demonstration to show how crowdsourcing significantly improves the performance of automatic type completion algorithms.

Attribute value matching (AVM) identifies equivalent values that refer to the same entities. Traditional approaches ignore the weights of values in itself. In this demonstration, we present AVM-LB, Attribute Value Matching with Limited Budget, that preferentially matches the hot equivalent values such that the maximal benefit to data consistency can be achieved by limited budget. By defining a rank function and greedily matching the hot equivalent values, the AVM-LB enables users to interactively explore the achieved benefit to data consistency.

ID-associated data are sequences of entries, and each entry is semantically associated with a unique ID. Examples are user IDs in user behaviour logs of mobile applications and device IDs in sensor records of self-driving cars. Nowadays, many big data applications generate such types of ID-associated data at high speed, and most queries over them are ID-centric (on specific IDs and ranges of time). To generate valuable insights from such data timely, the system needs to ingest high volumes of them with low latency, and support real-time analysis over them efficiently. In this paper, we introduce a system prototype designed for this goal. The system designed a parallel ingestion pipeline and a lightweight indexing scheme for the fast ingestion and efficient analysis. Besides, a fiber partitioning method is utilized to achieve dynamic scalability. For better integration with Hadoop ecosystem, the prototype is implemented based on open source projects, including Kafka and Presto.

In model driven Enterprise Information Systems (EIS) conceptual ER models are fundamental. We demonstrate how a phenomenological modeling ontology for conceptual models assists in creating meaningful, auto generated EIS. Application of the phenomenological ontology on conceptual modeling is shown in a graphical, easy-to-use CASE tool, CoreWEB, which support both editing of models and generating user interfaces for consistent manipulation of instance data. A visual and easy-to-grasp demonstration of conceptual modeling with live generation of EIS should be a welcome holistic complement to specialized, focused, in-depth presentations. Audience at all levels find food for thought and inspiration to continue to explore conceptual modeling in general and phenomena oriented approaches with systems generation in particular.

With the rapid growth of mashup resources, clustering mashup services according to the functions of the mashup services has become an effective way to improve the quality of mashup services management. Clustering is a learning task that classifies individuals or objects into different clusters based on the similarity. The purpose of clustering is to maximize the homogeneity of elements in the same cluster and maximize the heterogeneity of the elements in different clusters. It is a multivariate statistical method for classification. However, compared with the supervised classification, the clustering’s ability to categorize is much weaker. Existing methods for mashup services clustering mostly focus on utilizing key features from WSDL documents directly. In this paper, we proposed a method to improve the categorize ability of clustering. That is, applying supervised thought to cluster mashup services. First, taking basic clustering operations on the WSDL documents of mashups to obtain the clustering result for each element. Then, using the WSDL documents as training data, and the clustering results from the first step as pseudo-tags to train a classification learner. Finally, classifying mashups with this classification learner to get the final clustering results.

The temporal path problem is significant and challenging, where the connections between the vertices are temporal and there can be many attributes on the vertices and edges, such as vehicle speed and the price of a flight. Then in path finding, in addition to the single requirement of the length, or the arrival time, people would like to specify multiple constraints on the attributes to illustrate their requirements in real applications, such as the total cost, the total travel time and the stopover interval of a flight between two cities. In this paper, we devise a system called MCTP to answer the new popular Multi-Constrained Path Queries (MCPQs) in attributed temporal graphs. To the best of our knowledge, this is the first system that supports MCPQs.

This paper presents an intelligent criminal community detection system, called ACCDS, to support various criminal event detection tasks such as drug abuse behavior discovery and illegal pyramid selling organization detection, based on evolving social graphs. The system contains four main components: data collection, community social graph construction, criminal community detection and data visualization. First, the system collects a large amount of e-government data from several real communities. The raw data consist of demographic data, social relations, house visiting records, and sampled criminal records. To protect the privacy, we desensitize the real data using some data processing techniques, and extract the important features for profiling the human behaviors. Second, we use a large static social graph to model the social relations of all residents and a sequence of time-evolving graphs to model the house visiting data for each house owner. With the graph models, we formulate the criminal community detection tasks as the subgraph mining problem, and implement a subgraph detection algorithm based on frequent pattern mining. Finally, the system provides very user-friendly interfaces to visualize the detected results to the corresponding user.

Being able to model and analyze social dependencies is one of the most important feature of iStar modeling language. However, typical layout algorithms (e.g., hierarchical layout and circular layout) are not suitable for laying out iStar models, especially, strategic dependency models. As a result, iStar models are typically constructed and laid out in a manual way, which is time-consuming and tedious, especially when dealing with large-scale models. In this paper, we present a prototype tool which can automatically lay out iStar model in a meaningful way by taking into account the semantics of iStar model elements. In particular, we have developed and integrated an istarml parser with our tool, enabling automatic layout of iStar models that are specified in terms of istarml. We believe our tool would be particularly interesting for people who used to deal with large-scale iStar models.

When complex queries are executed in a database system, several files will be accessed such as relational tables, indexes and profiles. Monitoring database files enables us to better understand the query progress and the storage system. In this demo, we present a tool for monitoring file operations such as when the file is accessed and how many bytes are transferred for read and write. The access information is visually displayed at running time, and also stored as historical data for further analysis and optimization. We provide a graphical interface to report and display the results. MDBF is implemented inside a system kernel with a low overhead. The demonstration is performed by using real trajectory datasets and continuous queries employing three different indexes.

Finding a certain number of objects with optimum ranking based on spatial position constraints and given preference of attributes can be essential in numerous scenarios. In this paper, we demonstrate CusFinder, an interactive customer ranking query system to retrieve customers who favour a specific seller more than other people from the perspective of sellers, instead of retrieving sellers for a given customer similar to existing commercial systems. To make the query processing more efficient, a novel indexing is proposed to serve for query engine. Furthermore, we present the result of queries upon multiple visualization views with user-friendly interaction designs.

With data currency issues draw the attentions of both researchers and engineers, temporal data, which describes real world events with time tags in database, is playing a key role in data warehouse, data mining, and etc. At the same time, 4V features of big data give rise to the difficulties in comprehensive data quality management and data cleaning. On one hand, entity resolution methods are faced with challenges when dealing with temporal data. On another hand, multiple problems existing in data records are hard to be captured and repaired. Motivated by this, we address data quality evaluation and data cleaning issues in imprecise temporal data. This project aims to solve three key problems in temporal data quality improvement and cleaning: (1) Determining currency on imprecise temporal data, (2) Entity resolution on temporal data with incomplete timestamps, and (3) Data quality improvement on consistency and completeness with data currency. The purpose of this paper is to address the problem definitions and discuss the procedure framework and the solutions of improving the effectiveness of temporal data cleaning with multiple errors.

In the era of big data, people are dealing with data all the time. Data collection is the first step and foundation for many other downstream applications. Meanwhile, we observe that data collection is often entity-oriented, i.e., people usually collect data related to a specific entity. In most cases, people achieve entity-oriented data collection by manual query and filtering based on search engines or news applications. However, these methods are not very efficient and effective. In this paper, we consider designing reasonable process rules and integrating artificial intelligence algorithms to help people efficiently and effectively collect the target data related to the specific entity. Concretely, we propose an active workflow method to achieve this goal. The whole workflow method is composed of four processes: task modeling for data collection, Internet data collection, crowdsourcing data collection and multi-source data aggregation.

Although many methods have been developed in social signal processing (SSP) field during the last decade, several issues related to data management and scalability are still emerging. As the existing visual non-verbal behavior analysis (VNBA) systems are task-oriented, they do not have comprehensive data models, and they are biased towards particular data acquisition procedures, social cues and analysis methods. In this paper, we propose a data model for the visual non-verbal cues. The proposed model is privacy-preserving in the sense that it grants decoupling social cues extraction phase from analysis one. Furthermore, this decoupling allows to evaluate and perform different combinations of extraction and analysis methods. Apart from the decoupling, our model can facilitate heterogeneous data fusion from different modalities since it facilitates the retrieval of any combination of different modalities and provides deep insight into the relationships among the VNBA systems components.

Query/view synchronization upon the evolution of a database schema is a critical problem that has drawn the attention of many researchers in the database community. It entails rewriting queries and views to make them continue work on the new schema version. Although several techniques have been proposed for this problem, many issues need yet to be tackled for evolutions concerning the deletion of schema constructs, hence yielding loss of information. In this paper, we propose a new methodology to rewrite queries and views whose definitions are based on information that have been lost during the schema evolution process. The methodology exploits (relaxed) functional dependencies to automatically rewrite queries and views trying to preserve their semantics.

Knowledge graph embedding aims to embed both entities and relations into a low-dimensional space. Most existing methods of representation learning consider direct relations and some of them consider multiple-step relation paths. Although those methods achieve state-of-the-art performance, they are far from complete. In this paper, a noval path-augmented TransD (PTransD) model is proposed to improve the accuracy of knowledge graph embedding. This model uses two vectors to represent entities and relations. One of them represents the meaning of a(n) entity (relation), the other one is used to construct the dynamic mapping matrix. The PTransD model considers relation paths as translation between entities for representation learning. Experimental results on public dataset show that PTransD achieves significant and consistent improvements on knowledge graph completion.

Temporal constraints are an important aspect for the modeling of processes. We analyze the expressiveness of different features for the representation of temporally constrained processes. In particular we are able to show that the provision of non-contingency for activities and of references to start events of activities are redundant. We provide transformations for mapping process models employing the richer sets of features to process models using a reduced set of features. We show the equivalence of these process models and discuss the advantages of reducing redundancies in the representation of temporal constraints.

Double entry bookkeeping lies at the core of modern accounting. It is shaped by a fundamental conceptual pattern; a design decision that was popularised by Pacioli some 500 years ago and subsequently institutionalised into accounting practice and systems. Debits and credits are core components of this conceptual pattern. This paper suggests that a different conceptual pattern, one that does not have debits and credits as its components, may be more suited to some modern accounting information systems. It makes the case by looking at two conceptual design choices that permeate the Pacioli pattern; de se and directional terms - leading to a de se directional conceptual pattern. It suggests alternative design choices - de re and non-directional terms, leading to a de re non-directional conceptual pattern - have some advantages in modern complex, computer-based, business environments.

Nowadays, Information System (IS) security and Risk Management (RM) are required for every organization that wishes to survive in this networked and open world. Thus, more and more organizations tend to implement a security strategy based on an ISSRM (IS security RM) approach. However, the difficulty of dealing efficiently with ISSRM is currently growing, because of the complexity of current IS coming with the increasing number of risks organizations need to face. To use conceptual models to deal with RM issues, especially in the information security domain, is today an active research topic, and many modelling languages have been proposed in this way. However, a current challenge remains the cognitive effectiveness of the visual syntax of these languages, i.e. the effectiveness to convey information. Security risk managers are indeed not used to use modelling languages in their daily work, making this aspect of cognitive effectiveness a must-have for these modelling languages. Instead of starting defining a new cognitive effective modelling language, our objective is rather to assess and benchmark existing ones from the literature. The aim of this paper is thus to assess the cognitive effectiveness of CORAS, a modelling language focused on ISSRM.

Database design is traditionally taught as a process where requirements are captured in a conceptual model, then forward engineered into a logical model, followed by implementation in a physical model. Conceptual models are intended to be abstract and platform-independent, thus expressing aspects of the application being modeled without narrowing the design choices prematurely. The early stages of the process are more established in pedagogy, while the last stage, physical design, seems largely unexplored in the computing education literature. Moreover, if the choice for the logical model is relational, there are several possible implementation models; the design space widens again after the phase of transforming a conceptual model into a logical model. This paper explores teaching students about the design space for physical modeling. The contents of learning modules on physical design are presented, including scaffolding of technical content in an abstract (conceptual) manner, followed by connection to real-world analogues, culminating in a project that requires application of conceptual knowledge to explore the impact of physical design alternatives. The achievement of learning outcomes with and without the final project is assessed for courses taught in two consecutive academic terms. Future areas of research are discussed, such as expanding and refining the physical design space, the student preparation, and the types of impact investigated.

Context: Enterprise Architecture seeks to align organizational objectives with decisions associated to people, processes, information and technology. Different frameworks have been proposed for designing an enterprise architecture, such as TOGAF or Zachman. However, defining an enterprise architecture following these frameworks is not an easy task since it requires the alignment of organizational needs with technological decisions. Goal: This paper presents a methodological framework, called Simple Enterprise Architecture (SEA), that eases the definition of an enterprise architecture and provides concrete proof of the alignment between IT decisions and organizational needs. Method: The SEA framework has been developed by integrating components from existing ones in order to generate a concrete process that guides analysts in the correct definition of an enterprise architecture. Results: This framework has been used in teaching system architecture master courses for 5 years with positive results.

Process modelling notations fall in two broad categories: declarative notations, which specify the rules governing a process; and imperative notations, which specify the flows admitted by a process. We outline an empirical approach to addressing the question of whether certain process logs are better suited for mining to imperative than declarative notations. We plan to attack this question by applying a flagship imperative and declarative miner to a standard collection of process logs, then evaluate the quality of the output models w.r.t. the standard model metrics of precision and generalisation. This approach requires perfect fitness of the output model, which substantially narrows the field of available miners; possible candidates include Inductive Miner and MINERful. With the metrics in hand, we propose to statistically evaluate the hypotheses that (1) one miner consistently outperforms the other on one of the metrics, and (2) there exist subsets of logs more suitable for imperative respectively declarative mining.

Experimental research in conceptual modeling typically involves comparing grammars or variations within a grammar, where differences between experimental groups are based on a focal construct of interest. However, a conceptual modeling grammar is a collection of many constructs and there is a danger that grammatical features other than those under consideration in an experiment can influence or confound the results obtained. To address this issue, we propose the use of artifact sampling as a way to systematically vary non-focal grammatical features in experimental conceptual modeling research to control for potential confounds or interactions between constructs of interest and other grammatical features. In this paper, we describe the approach and illustrate its application to the design of a large-scale study to compare alternative notations within the Entity-Relationship family of grammars.

Generating meaningful layout of iStar models is a challenging task, which currently requires significant manual efforts. However, it is time-consuming when dealing with large-scale iStar modeling, rising the need of having an automatic iStar layout tool. Previously, we have proposed an algorithm for laying out iStar SD models and have implemented a corresponding prototype tool. In this paper, we report our ongoing empirical work which aims to evaluate the effectiveness and usability of the prototype tool. In particular, we present a research design which is applied to compare manual layout and automatic layout in terms of efficiency and model comprehensibility. Based on such a design, we are planning to carry out empirical studies accordingly in the near future.

The notion of capability is used by practitioners and researchers alike to enable better understanding of business trajectories and the role of IT in achieving them. Building on the origins of the concept from strategic management, this paper lays out the requirements for capturing enterprise-specific and social characteristics of capabilities. The paper proposes adoption of a goal-driven agent-oriented modeling approach to satisfy the requirements. The ability of such an approach to explicate social and technical design alternatives and enable decision making on their tradeoffs is illustrated on a BPM capability.

Business process compliance (BPC) denotes business processes that adhere to requirements originating from different sources, e.g., laws or regulations. Compliance measures are used in business processes to prevent compliance violations and their consequences, such as fines or monetary sanctions. Compliance measures also incur costs, e.g., for tools, hardware, or personnel. To ensure that companies can work economically even in intensively regulated environments, the economic viability of BPC has to be taken into account. A body of literature is already devoted to the economic assessment of processes and focuses on the business perspective, whereas corresponding approaches for BPC appear to be lacking. Consequently, we introduce a novel approach that allows for an economic assessment of process-based compliance measures. The approach takes monetary consequences of compliance violations into account and is based on the well-known basic workflow patterns for control flows. We demonstrate its applicability by means of an exemplary ordering process affected by Article 32 (1) of the EU General Data Protection Regulation.

TOGAF Architecture Development Method (ADM) covers different aspects of Enterprise Architecture (EA) management as well as it provides textual guidelines to adapt and perform EA processes including the requirements management (RM) process. We observed that adopting ADM following these guidelines is an intricate task because the effort required to define a sequential interaction between related activities is meticulous and hard. During a real case experience we have formalized the ADM-TOGAF Requirements Management textual guidelines with the Business Process Model and Notation (BPMN) to facilitate its comprehension and usage. Afterwards, the usefulness of the proposed process models has been qualified with a group of engineering and master students.

In this paper, we describe a requirements engineering method with a focus on the conceptualization of alternative service offerings. The practical context for our project is based on the first author’s work in a startup. Our proposed method is suitable for exploring market opportunities while specifying a service offering. Our method helps requirements engineering practitioners understand the business and technology worlds by modeling business needs and technical capabilities in the same model.

Pivoting is used by many startups and large enterprises to reconfigure their structures and relationships in line with their changing environments and requirements. However, pivoting is a non-trivial undertaking that has far reaching consequences for the focal organization. Conceptual models of actor intentionality can be used to design and analyze organizational pivots in a systematic and structured manner. Conceptual modeling is preferable to ad hoc evaluation as it can provide more detailed and systematic analysis of pivoting decisions. It can be used to uncover mistakes and gaps in reasoning that are missed or obscured via ad hoc evaluation. Actor- and goal-modeling can be used to differentiate among beneficial and deleterious pivoting options. Correctly designed and implemented pivots can avoid substantial value destruction from direct damages as well as opportunity loss for the focal organization. In this paper we present conceptual models of pivoting based on a retrospective case example of Twitter.

Management of big spatio-temporal data such as the results from large scale global climate models has long been a challenge because of the sheer vastness of the dataset. Although different data management systems like that incorporate a relational database management system have been proposed and widely used in prior studies, solutions that are particularly designed for big spatio-temporal data management have not been studied well. In this paper, we propose a general data management platform for high-dimensional spatio-temporal datasets like those found in the climate domain, where different database systems can be applied. Through this platform, we compare and evaluate several database systems including SQL database and NoSQL database from various aspects and explore the key impact factors for system performance. Our experimental results indicate advantages and disadvantages of each database system and give insight into the best system to use for big spatio-temporal data applications. Our analysis provides important insights into the understanding of performance of different data management systems, which is very useful for designing high dimensional big data applications.

One of the greatest challenges in big data management is the “Variety” of the data. The data may be presented in various types and formats: structured, semi-structured and unstructured. For instance, data can be modeled as relational, key-value, and graph models. Having a single data platform for managing both well-structured data and NoSQL data is beneficial to users; this approach reduces significantly integration, migration, development, maintenance, and operational issues. Therefore, a challenging research work is how to develop an efficient consolidated single data management platform covering both NoSQL and relational data to reduce integration issues, simplify operations, and eliminate migration issues. In this paper, we envision novel principles and technologies to handle multiple models of data in one unified database system, including model-agnostic storage, unified query processing and indexes, in-memory structures and multi-model transactions. We discuss our visions as well as present research challenges that we need to address.

This paper develops a model of conflicts that relies on extracting text and argument features from traces of interactions in collaborative work. Much prior research about collaborative work is aimed at improving the support for virtual work. In contrast, we are interested in detecting conflicts in collaborative work because conflict undetected can escalate and cause disruptions to productive work. It is a difficult problem because it requires untangling conflict-related interactions from normal interactions. Few models or methods are available for this purpose. The extracted features, interpreted with the help of foundational theories, suggests a conceptual model of conflicts that include categories of argumentation such as reasoning and modality; and informative language features. We illustrate the extraction approach and the model with a dataset from Bugzilla. The paper concludes with a discussion of evaluation possibilities and potential implications of the approach for detecting and managing conflicts in collaborative work.

Data quality management in document oriented data stores has not been deeply explored yet, presenting many challenges that arise because of the lack of a rigid schema associated to data. Data quality is a critical aspect in this kind of data stores, since its control is not possible and it is not a priority in the data storage stage. Additionally, data quality evaluation and improvement are also very difficult tasks due to the schema-less characteristic of data. This paper presents a first step towards data quality management in document oriented data stores. In order to address the problem, the paper proposes a strategy for defining data granularities for data quality evaluation and analyses some data quality dimensions relevant to document stores.

If there is a domain where data management becomes an intensive Big Data issue, it is the genomic domain, due to the fact that the data generated day after day are exponentially increasing. A genomic data management strategy requires the use of a systematic method, intended to assure that the right data are identified, using the adequate data sources, and linking the selected information with a software platform based on conceptual models, which allows guaranteeing the implementation of genomic services with quality, efficient and valuable data. In this paper, we select the method called “SILE” –for Search, Identification, Load and Exploitation-, and we focus on validating its accuracy in the context of a concrete disease, the Colorectal Cancer. The main contribution of our work is to show how such methodological approach can be applied successfully in a real and complex clinical context, providing a working environment where Genomic Big Data are efficiently managed.

Digital Ecosystems consist of a variety of interlinked subsystems. This paper presents a flexible approach to define the links between such subsystems. The idea is to exploit the paradigm of Model Centered Architecture (MCA) and to specify all links/interfaces by means of appropriate Domain Specific Modeling Languages. The approach has been successfully applied and evaluated in several projects. As a proof of concept, we present the model-based interfacing between assistive systems and human activity recognition systems, which showed good performance as needed in real-world applications.