Lei Shi | University of Bath (original) (raw)
Papers by Lei Shi
An engineering service project can be highly interactive, collaborative and distributed. The impl... more An engineering service project can be highly interactive, collaborative and distributed. The implementation of such projects needs to generate, utilise and share large amounts of data and heterogeneous digital objects. The information overload prevents the effective reuse of project data and knowledge, makes the understanding of project characteristics to become difficult. Towards solving these issues, this paper emphasised the using of data mining and machine learning techniques to improve the project characteristic understanding process. The work presented in this paper proposed an automatic model and some analytical approaches for learning and predicting the characteristics of engineering service projects. To evaluate the model and demonstrate its functionalities, an industrial dataset from the aerospace sector is considered by the case study. This work shows that the proposed model could enable the project members to gain comprehensive understanding of project characteristics from a multi-dimensional perspective, and it has the potential to support them in implementing evidence-based design and decision-making.
Process management is considered as an essential approach to improve the performance of an enterp... more Process management is considered as an essential approach to improve the performance of an enterprise. The process of an engineering project can be considered to be a formalised workflow accompanied by a set of decisions. With decisions being made through taking account of information from various sources, the operation and management of modern engineering projects need to deal with increasing amounts of dynamic and changing information. Hence, understanding this dynamic project information and interpreting the information for process management can generate challenges in practice. This might be caused by constraints of time and resource, the distributed structure of the information and a lack of modelled domain knowledge. To address these challenges, this paper focuses on techniques generated for supporting the automation of the process management of engineering projects from a data-driven perspective. It includes elements of process modelling, process monitoring and process evaluation of such projects, through a proposed system framework, working as the data evolves over time. The design and implementation of the system framework and related approaches apply autonomic computing, data mining and knowledge management technologies. To demonstrate the system functionality, a dataset from an aerospace organisation is considered and described in the paper.
Product service plays an essential role in day-to-day operations of nowadays manufacturing indust... more Product service plays an essential role in day-to-day operations of nowadays manufacturing industries. However, the changing demands of the market/customers, the increasing complexity of product functionalities and the extended product lifecycles present challenges to related In-Service projects. In order to handle the increasing number of projects and to control the costs and resource consumptions, it is critical to improve the efficiency and automation of process management. Within this context, this paper introduces some datadriven approaches to interpret and represent changes of project process over time in an automatic manner. These approaches aim to help project actors improve their understanding of process structure and the efficiency of process management, and also enable them to investigate process changes from more dynamic perspectives. To evaluate the approaches, a dataset from an aerospace organisation is considered in this paper.
International Conference on Manufacturing Research, Sep 2015
Many modern engineering projects are required to coordinate multiple project teams, utilise dist... more Many modern engineering projects are required to coordinate multiple project teams, utilise
distributed resources and integrate knowledge across multiple disciplines. Hence, the execution of
these projects needs to involve high volume of remote communications, synchronous or asynchronous
interactions, complicated decision-making and control processes. By considering these factors,
human-centred management approaches could be problematic at times to handle the large amount of
project data and manage the complex project processes. In order to improve the management
efficiency and effectiveness, this paper presents an automatic approach on process reconstruction and
process visualisation. As shown in a case study, this approach has the potential to support the
information reuse, capture process dynamics, enhance process comprehensibility, as well as reduce
human intervention in general process management.
International Conference on Manufacturing Research, Sep 2015
Open Design and crowd-sourced development have recently formed a new paradigm in the engineering ... more Open Design and crowd-sourced development have recently formed a new paradigm in the engineering design and manufacture industries. Through collaboration and free sharing of information, open design has proven capable of developing complex products at a low cost, utilising an expert and motivated work force to bring products to market. However, as open design differs greatly from traditional engineering enterprise in several ways, there are challenges in its effective utilisation. This paper proposes an approach (termed engineering project health monitoring) to addressing two challenges; automatically generating a shared understanding of project-specific information, and providing the means to aid workflow management in the open design situation. As open design is often performed through distributed individual working, and is often subject to little management structure, aiding the generation of a shared understanding and control of activity scheduling and prioritisation are vital to ensuring a high-performing output.
DESIGN 2014, May 23, 2014
The generation of digital objects are present throughout the work that is undertaken during the e... more The generation of digital objects are present throughout the work that is undertaken during the engineering design process. Although it is self-evident that these objects have been created by engineers to support their work, little research has been performed to understand whether there are underlying signatures in their generation that could provide useful insights to the management of the project. In order to address this, this paper presents some initial insights from the analysis of the generation of digital objects from an engineering project, and discusses potential signatures.
The generation of digital objects are present throughout the work that is undertaken during the e... more The generation of digital objects are present throughout the work that is undertaken during the engineering design process. Although it is self-evident that these objects have been created by engineers to support their work, little research has been performed to understand whether there are underlying signatures in their generation that could provide useful insights to the management of the project. In order to address this, this paper presents some initial insights from the analysis of the generation of digital objects from an engineering project, and discusses potential signatures.
DESIGN 2014: International Design Conference. Dubrovnik, Croatia
Engineering projects are highly complex, involving many project actors as they produce thousands ... more Engineering projects are highly complex, involving many project actors as they produce thousands of files, emails and prototypes. The premise of this work is that in these complex project ecosystems there is a high amount of implied information, particularly in the digital files produced by each actor, which has potential to be automatically extracted and used to improve process or output. This work presents a framework by which files can be categorised and analysed, thereby generating useful project information directly from the actions of project actors; and gives two real-data examples.
This paper presents a topic-based analysis of email subject line data from a large-scale engineer... more This paper presents a topic-based analysis of email subject line data from a large-scale engineering project and explores its utility for supporting the management of collaborative work. The main contributions of the paper are a novel interpretation of the co-word network analysis method for application within an engineering project management context, and the appraisal of the method for finding patterns within subject line data. Our findings suggest that the approach has the potential to contribute to monitoring work complexity, tracking progress, recognizing synergy and divergence, detecting scope creep, and supporting knowledge capture.
Nowadays manufacturing involves high volume of complex operational processes, distributed resourc... more Nowadays manufacturing involves high volume of complex operational processes, distributed resources and international/intersectional collaborations, which cause the evaluation of performance for related engineering projects to become a challenge. The performance of project is a key factor that determines the quality of output, and its temporal changes could reflect the status of project execution at current time, as well as indicate the potential issues for the near future. As a result, monitoring the change of performance is an essential approach to ensure the project execution is on track. It could raise awareness of project participants upon any issue occurs, and enable them to make appropriate decisions on a real-time basis.
To facilitate the evaluation of project performance in collaborative environments, this research aims to propose an automatic approach to extracting key performance indicators (KPIs) from project related data, and also to demonstrate how the domain knowledge can facilitate the process of KPIs identification and visualisation.
Communication features in many engineering activities within an engineering project. It is the ma... more Communication features in many engineering activities within an engineering project. It is the main form by which information & knowledge is shared, and facilitates the generation of a shared understanding between engineers. While there exists a significant body of literature relating to communication, much of the research has been through qualitative studies using techniques such as surveys, interviews and observation. Given the prevalence of computer-mediated communication and the development of techniques to analyse such datasets, there is now the opportunity to provide quantitative metrics that can characterise communication.
Therefore, this paper examines this opportunity through the co-word analysis of the subject line terms of an engineering project e-mail corpus comprising of 10,628 e-mails, featuring 1,045 individuals and spanning over a 4 year period. More specifically, the analysis has focused on the evolution, use/re-use and centrality of terms across the various project stages. The results provide interesting insights in the evolution of engineering terminology, which leads onto a discussion on how these metrics may provide indicators of project 'normality'.
Modern engineering projects are often large, complex, high-value, high-risk, and distributed. As ... more Modern engineering projects are often large, complex, high-value, high-risk, and distributed. As a result, it is both vital to monitor and understand what is happening within each as it progresses, and highly challenging to do so. Without detailed and comprehensive understanding, management becomes difficult and falls back upon generic high-level principles that are not always appropriate for each project context. To approach this issue, this paper studies the written logbooks of three engineers, and explores how the marks within can be analysed to generate project-level understanding, particularly that which can inform engineering project management. This occurs through the study of three engineering logbooks using two detailed coding schemas, one classifying content and the other activity, creativity and novelty. By this analysis, this paper aims to understand and assess efficacy of studying logbooks given their time-consuming and difficult-to-code nature. From the results, feasibility is demonstrated of developing detailed understanding of typical project progress, and the identification of specific events within a project upon which a manager may act. The efficacy of the study of logbooks for this purpose is then assessed.
DESIGN 2014: International Conference on Engineering Design. Dubrovnik, Croatia, May 20, 2014
Tools with Artificial Intelligence (ICTAI), 2014 IEEE 26th International Conference on
Engineering projects are often highly complex, unique and safety critical, which can lead to the ... more Engineering projects are often highly complex, unique and safety critical, which can lead to the complex engineering processes requiring various resources and bring challenges on decision making. To ensure the success of the project products, the related project has to comply with stringent regulations and company processes before the work can commence. In addition, the increasing in-service lifespan of products has led to many projects, such as design or maintenance projects, being run concurrently in a highly time-constrained and high-pressured environment, so that the generation of detailed activity plans becomes a challenge for most engineering companies. These plans are typically achieved through the ability of the project managers to use their knowledge, experience and constant contact with the engineers to provide insights into the sequence of engineering activity. However, the viability of the current method to manually generate and evaluate the activity plan is becoming an issue, e.g. there are increasing numbers of concurrent projects with increasingly distributed resources and higher level of complexity.
As regulatory and/or company process demands, the project related data is archived within this context and thus, provides a wealth of potentially useful information that could be utilised in the management of current projects. Therefore, this research investigates the potential value provided by the automatic construction of past project activity sequences, and proposes analytical methods to represent the normality of the projects based on the extracted patterns from their sequences. The evaluation applies industrial data, and shows that the results generated by the proposed approach can accurately reflect the process similarity and normality of the projects.
3rd Annual EPSRC Manufacturing the Future Conference
Knowledge is consistently generated and used across the manufacturing project lifecycle. To preve... more Knowledge is consistently generated and used across the manufacturing project lifecycle. To prevent obsolescence and improve sustainability, certain project knowledge needs to be preserved and re-used. The proposed approach is an automatic mechanism and requires zero user intervention, thus it can be smoothly applied or integrated in real manufacturing environments. It can discover the knowledge from “big data”, and then uses the prediction mechanism to provide various assistants to project participants.
DESIGN 2014: International Conference on Engineering Design. Dubrovnik, Croatia
The increasing complexity of current engineering projects causes uncertainty for project executio... more The increasing complexity of current engineering projects causes uncertainty for project execution, and creates challenges for project design and management. To attempt to reuse the information of previously completed projects for current project planning, this paper introduces the definition of the “Engineering Document Pattern” (EDP), and proposes some EDP identification methods. EDP is really a form of knowledge and activity representation and as an approach is suitable for representing large scale projects, and can potentially be used to help decision makers understand the initial requirements and potential risks of current projects. The experimental study evaluates the EDP identification methods by analysing and profile some completed In-Service design repair reports, and the result shows the identified EDPs accurately represent the report-related project characteristics.
DESIGN 2014: International Conference on Engineering Design. Dubrovnik, Croatia
Engineering projects are highly complex, involving many project actors as they produce thousands ... more Engineering projects are highly complex, involving many project actors as they produce thousands of files, e-mails and outputs. The premise of this work is that in these complex project eco-systems there is a high amount of implied information, particularly in the digital files produced by each actor, which has potential to be automatically extracted and used to improve process or output. This work presents a framework by which files can be categorised and analysed, thereby generating useful project information directly from the actions of project actors; and gives two real-data examples.
DESIGN 2014: International Conference on Engineering Design. Dubrovnik, Croatia
The generation of digital objects are present throughout the work that is undertaken during the e... more The generation of digital objects are present throughout the work that is undertaken during the engineering design process. Although it is self-evident that these objects have been created by engineers to support their work, little research has been performed to understand whether there are underlying signatures in their generation that could provide useful insights to the management of the project. In order to address this, this paper presents some initial insights from the analysis of the generation of digital objects from an engineering project, and discusses potential signatures.
International Journal of Knowledge-Based and Intelligent Engineering Systems
This research addresses the semantic and knowledge gap problem in information retrieval by propos... more This research addresses the semantic and knowledge gap problem in information retrieval by proposing an ontologybased semantic feature-matching approach, which uses natural language processing, named entity recognition and user-oriented ontologies. The approach comprises four steps: (i) user-oriented ontology building; (ii) semantic feature extraction for identifying information objects; (iii) semantic feature selection using user-oriented ontologies to enhance the semantic representation of the information objects, and (iv) measuring the similarity between the information objects using their enhanced semantic representations. The experiment conducted explores the retrieval performance of the proposed approach and shows that it consistently outperforms its corresponding term-based approach by demonstrating improved precision, recall and F-score.
In: Frontiers in Artificial Intelligence and Applications. IOS Press, pp. 1887-1898.
An engineering service project can be highly interactive, collaborative and distributed. The impl... more An engineering service project can be highly interactive, collaborative and distributed. The implementation of such projects needs to generate, utilise and share large amounts of data and heterogeneous digital objects. The information overload prevents the effective reuse of project data and knowledge, makes the understanding of project characteristics to become difficult. Towards solving these issues, this paper emphasised the using of data mining and machine learning techniques to improve the project characteristic understanding process. The work presented in this paper proposed an automatic model and some analytical approaches for learning and predicting the characteristics of engineering service projects. To evaluate the model and demonstrate its functionalities, an industrial dataset from the aerospace sector is considered by the case study. This work shows that the proposed model could enable the project members to gain comprehensive understanding of project characteristics from a multi-dimensional perspective, and it has the potential to support them in implementing evidence-based design and decision-making.
Process management is considered as an essential approach to improve the performance of an enterp... more Process management is considered as an essential approach to improve the performance of an enterprise. The process of an engineering project can be considered to be a formalised workflow accompanied by a set of decisions. With decisions being made through taking account of information from various sources, the operation and management of modern engineering projects need to deal with increasing amounts of dynamic and changing information. Hence, understanding this dynamic project information and interpreting the information for process management can generate challenges in practice. This might be caused by constraints of time and resource, the distributed structure of the information and a lack of modelled domain knowledge. To address these challenges, this paper focuses on techniques generated for supporting the automation of the process management of engineering projects from a data-driven perspective. It includes elements of process modelling, process monitoring and process evaluation of such projects, through a proposed system framework, working as the data evolves over time. The design and implementation of the system framework and related approaches apply autonomic computing, data mining and knowledge management technologies. To demonstrate the system functionality, a dataset from an aerospace organisation is considered and described in the paper.
Product service plays an essential role in day-to-day operations of nowadays manufacturing indust... more Product service plays an essential role in day-to-day operations of nowadays manufacturing industries. However, the changing demands of the market/customers, the increasing complexity of product functionalities and the extended product lifecycles present challenges to related In-Service projects. In order to handle the increasing number of projects and to control the costs and resource consumptions, it is critical to improve the efficiency and automation of process management. Within this context, this paper introduces some datadriven approaches to interpret and represent changes of project process over time in an automatic manner. These approaches aim to help project actors improve their understanding of process structure and the efficiency of process management, and also enable them to investigate process changes from more dynamic perspectives. To evaluate the approaches, a dataset from an aerospace organisation is considered in this paper.
International Conference on Manufacturing Research, Sep 2015
Many modern engineering projects are required to coordinate multiple project teams, utilise dist... more Many modern engineering projects are required to coordinate multiple project teams, utilise
distributed resources and integrate knowledge across multiple disciplines. Hence, the execution of
these projects needs to involve high volume of remote communications, synchronous or asynchronous
interactions, complicated decision-making and control processes. By considering these factors,
human-centred management approaches could be problematic at times to handle the large amount of
project data and manage the complex project processes. In order to improve the management
efficiency and effectiveness, this paper presents an automatic approach on process reconstruction and
process visualisation. As shown in a case study, this approach has the potential to support the
information reuse, capture process dynamics, enhance process comprehensibility, as well as reduce
human intervention in general process management.
International Conference on Manufacturing Research, Sep 2015
Open Design and crowd-sourced development have recently formed a new paradigm in the engineering ... more Open Design and crowd-sourced development have recently formed a new paradigm in the engineering design and manufacture industries. Through collaboration and free sharing of information, open design has proven capable of developing complex products at a low cost, utilising an expert and motivated work force to bring products to market. However, as open design differs greatly from traditional engineering enterprise in several ways, there are challenges in its effective utilisation. This paper proposes an approach (termed engineering project health monitoring) to addressing two challenges; automatically generating a shared understanding of project-specific information, and providing the means to aid workflow management in the open design situation. As open design is often performed through distributed individual working, and is often subject to little management structure, aiding the generation of a shared understanding and control of activity scheduling and prioritisation are vital to ensuring a high-performing output.
DESIGN 2014, May 23, 2014
The generation of digital objects are present throughout the work that is undertaken during the e... more The generation of digital objects are present throughout the work that is undertaken during the engineering design process. Although it is self-evident that these objects have been created by engineers to support their work, little research has been performed to understand whether there are underlying signatures in their generation that could provide useful insights to the management of the project. In order to address this, this paper presents some initial insights from the analysis of the generation of digital objects from an engineering project, and discusses potential signatures.
The generation of digital objects are present throughout the work that is undertaken during the e... more The generation of digital objects are present throughout the work that is undertaken during the engineering design process. Although it is self-evident that these objects have been created by engineers to support their work, little research has been performed to understand whether there are underlying signatures in their generation that could provide useful insights to the management of the project. In order to address this, this paper presents some initial insights from the analysis of the generation of digital objects from an engineering project, and discusses potential signatures.
DESIGN 2014: International Design Conference. Dubrovnik, Croatia
Engineering projects are highly complex, involving many project actors as they produce thousands ... more Engineering projects are highly complex, involving many project actors as they produce thousands of files, emails and prototypes. The premise of this work is that in these complex project ecosystems there is a high amount of implied information, particularly in the digital files produced by each actor, which has potential to be automatically extracted and used to improve process or output. This work presents a framework by which files can be categorised and analysed, thereby generating useful project information directly from the actions of project actors; and gives two real-data examples.
This paper presents a topic-based analysis of email subject line data from a large-scale engineer... more This paper presents a topic-based analysis of email subject line data from a large-scale engineering project and explores its utility for supporting the management of collaborative work. The main contributions of the paper are a novel interpretation of the co-word network analysis method for application within an engineering project management context, and the appraisal of the method for finding patterns within subject line data. Our findings suggest that the approach has the potential to contribute to monitoring work complexity, tracking progress, recognizing synergy and divergence, detecting scope creep, and supporting knowledge capture.
Nowadays manufacturing involves high volume of complex operational processes, distributed resourc... more Nowadays manufacturing involves high volume of complex operational processes, distributed resources and international/intersectional collaborations, which cause the evaluation of performance for related engineering projects to become a challenge. The performance of project is a key factor that determines the quality of output, and its temporal changes could reflect the status of project execution at current time, as well as indicate the potential issues for the near future. As a result, monitoring the change of performance is an essential approach to ensure the project execution is on track. It could raise awareness of project participants upon any issue occurs, and enable them to make appropriate decisions on a real-time basis.
To facilitate the evaluation of project performance in collaborative environments, this research aims to propose an automatic approach to extracting key performance indicators (KPIs) from project related data, and also to demonstrate how the domain knowledge can facilitate the process of KPIs identification and visualisation.
Communication features in many engineering activities within an engineering project. It is the ma... more Communication features in many engineering activities within an engineering project. It is the main form by which information & knowledge is shared, and facilitates the generation of a shared understanding between engineers. While there exists a significant body of literature relating to communication, much of the research has been through qualitative studies using techniques such as surveys, interviews and observation. Given the prevalence of computer-mediated communication and the development of techniques to analyse such datasets, there is now the opportunity to provide quantitative metrics that can characterise communication.
Therefore, this paper examines this opportunity through the co-word analysis of the subject line terms of an engineering project e-mail corpus comprising of 10,628 e-mails, featuring 1,045 individuals and spanning over a 4 year period. More specifically, the analysis has focused on the evolution, use/re-use and centrality of terms across the various project stages. The results provide interesting insights in the evolution of engineering terminology, which leads onto a discussion on how these metrics may provide indicators of project 'normality'.
Modern engineering projects are often large, complex, high-value, high-risk, and distributed. As ... more Modern engineering projects are often large, complex, high-value, high-risk, and distributed. As a result, it is both vital to monitor and understand what is happening within each as it progresses, and highly challenging to do so. Without detailed and comprehensive understanding, management becomes difficult and falls back upon generic high-level principles that are not always appropriate for each project context. To approach this issue, this paper studies the written logbooks of three engineers, and explores how the marks within can be analysed to generate project-level understanding, particularly that which can inform engineering project management. This occurs through the study of three engineering logbooks using two detailed coding schemas, one classifying content and the other activity, creativity and novelty. By this analysis, this paper aims to understand and assess efficacy of studying logbooks given their time-consuming and difficult-to-code nature. From the results, feasibility is demonstrated of developing detailed understanding of typical project progress, and the identification of specific events within a project upon which a manager may act. The efficacy of the study of logbooks for this purpose is then assessed.
DESIGN 2014: International Conference on Engineering Design. Dubrovnik, Croatia, May 20, 2014
Tools with Artificial Intelligence (ICTAI), 2014 IEEE 26th International Conference on
Engineering projects are often highly complex, unique and safety critical, which can lead to the ... more Engineering projects are often highly complex, unique and safety critical, which can lead to the complex engineering processes requiring various resources and bring challenges on decision making. To ensure the success of the project products, the related project has to comply with stringent regulations and company processes before the work can commence. In addition, the increasing in-service lifespan of products has led to many projects, such as design or maintenance projects, being run concurrently in a highly time-constrained and high-pressured environment, so that the generation of detailed activity plans becomes a challenge for most engineering companies. These plans are typically achieved through the ability of the project managers to use their knowledge, experience and constant contact with the engineers to provide insights into the sequence of engineering activity. However, the viability of the current method to manually generate and evaluate the activity plan is becoming an issue, e.g. there are increasing numbers of concurrent projects with increasingly distributed resources and higher level of complexity.
As regulatory and/or company process demands, the project related data is archived within this context and thus, provides a wealth of potentially useful information that could be utilised in the management of current projects. Therefore, this research investigates the potential value provided by the automatic construction of past project activity sequences, and proposes analytical methods to represent the normality of the projects based on the extracted patterns from their sequences. The evaluation applies industrial data, and shows that the results generated by the proposed approach can accurately reflect the process similarity and normality of the projects.
3rd Annual EPSRC Manufacturing the Future Conference
Knowledge is consistently generated and used across the manufacturing project lifecycle. To preve... more Knowledge is consistently generated and used across the manufacturing project lifecycle. To prevent obsolescence and improve sustainability, certain project knowledge needs to be preserved and re-used. The proposed approach is an automatic mechanism and requires zero user intervention, thus it can be smoothly applied or integrated in real manufacturing environments. It can discover the knowledge from “big data”, and then uses the prediction mechanism to provide various assistants to project participants.
DESIGN 2014: International Conference on Engineering Design. Dubrovnik, Croatia
The increasing complexity of current engineering projects causes uncertainty for project executio... more The increasing complexity of current engineering projects causes uncertainty for project execution, and creates challenges for project design and management. To attempt to reuse the information of previously completed projects for current project planning, this paper introduces the definition of the “Engineering Document Pattern” (EDP), and proposes some EDP identification methods. EDP is really a form of knowledge and activity representation and as an approach is suitable for representing large scale projects, and can potentially be used to help decision makers understand the initial requirements and potential risks of current projects. The experimental study evaluates the EDP identification methods by analysing and profile some completed In-Service design repair reports, and the result shows the identified EDPs accurately represent the report-related project characteristics.
DESIGN 2014: International Conference on Engineering Design. Dubrovnik, Croatia
Engineering projects are highly complex, involving many project actors as they produce thousands ... more Engineering projects are highly complex, involving many project actors as they produce thousands of files, e-mails and outputs. The premise of this work is that in these complex project eco-systems there is a high amount of implied information, particularly in the digital files produced by each actor, which has potential to be automatically extracted and used to improve process or output. This work presents a framework by which files can be categorised and analysed, thereby generating useful project information directly from the actions of project actors; and gives two real-data examples.
DESIGN 2014: International Conference on Engineering Design. Dubrovnik, Croatia
The generation of digital objects are present throughout the work that is undertaken during the e... more The generation of digital objects are present throughout the work that is undertaken during the engineering design process. Although it is self-evident that these objects have been created by engineers to support their work, little research has been performed to understand whether there are underlying signatures in their generation that could provide useful insights to the management of the project. In order to address this, this paper presents some initial insights from the analysis of the generation of digital objects from an engineering project, and discusses potential signatures.
International Journal of Knowledge-Based and Intelligent Engineering Systems
This research addresses the semantic and knowledge gap problem in information retrieval by propos... more This research addresses the semantic and knowledge gap problem in information retrieval by proposing an ontologybased semantic feature-matching approach, which uses natural language processing, named entity recognition and user-oriented ontologies. The approach comprises four steps: (i) user-oriented ontology building; (ii) semantic feature extraction for identifying information objects; (iii) semantic feature selection using user-oriented ontologies to enhance the semantic representation of the information objects, and (iv) measuring the similarity between the information objects using their enhanced semantic representations. The experiment conducted explores the retrieval performance of the proposed approach and shows that it consistently outperforms its corresponding term-based approach by demonstrating improved precision, recall and F-score.
In: Frontiers in Artificial Intelligence and Applications. IOS Press, pp. 1887-1898.
The increasing complexity of current engineering projects causes uncertainty for project executio... more The increasing complexity of current engineering projects causes uncertainty for project execution, and creates challenges for project design and management. To attempt to reuse the information of previously completed projects for current project planning, this paper introduces the definition of the “Engineering Document Pattern” (EDP), and proposes some EDP identification methods. EDP is really a form of knowledge and activity representation and as an approach is suitable for representing large scale projects, and can potentially be used to help decision makers understand the initial requirements and potential risks of current projects. The experimental study evaluates the EDP identification methods by analysing and profile some completed In-Service design repair reports, and the result shows the identified EDPs accurately represent the report-related project characteristics.
International Conference on Product Lifecycle Management (PLM15), Oct 2015
The development of Computer Aided Design (CAD) models is a fundamental and distinct feature of En... more The development of Computer Aided Design (CAD) models is a fundamental and distinct feature of Engineering Projects. CAD models can be considered to be the digital embodiment of the products’ design and are used to support a wide variety of tasks that span the embodiment, detail, manufacture and commissioning phases of a project. With this in mind, it is proposed that the monitoring and modelling of the edit trace behaviour of CAD files may provide additional understanding and evidence that supplements current approaches to monitor and manage engineering projects.
To explore this proposition, this paper presents results from an exploratory study that seeks to model the edit trace behaviour of CAD files based upon their meta- data attributes (for example, file size, date modified & date accessed). The edit trace behaviour has been mapped to a sigmoid function in order to be able to describe and potentially predict future behaviour. The potential impact of being able to provide this type of information to engineering project management is also discussed.
Computer Aided Design (CAD) has become an integral tool for many engineering design activities. C... more Computer Aided Design (CAD) has become an integral tool for many engineering design activities. CAD not only forms the digital embodiment of the products design but also supports communication between engineers as a intermediary object, the analysis of engineering systems, and enables computer-mediated collaborative work on the product. It is argued that the logs of the engineers' user interactions could be a powerful sensor in providing insights into the engineers and engineering design process. This paper presents results from the analysis of CAD logs generated by 45 engineering students.