A semantic search system for efficient information retrieval in the construction domain

Purpose – This research aims to improve the efficiency of information retrieval in a construction or renovation project, the aim is to deliver the right information to the right stakeholder at the right time throughout the project lifecycle. Design/methodology/approach – The approach involves creating a knowledge graph-based semantic search system. This system is structured by developing an ontology to organize the knowledge graph, which will adhere to linked data principles. Three key indicators—relevance, contextual awareness and integrity—are developed to assess query results, ensuring that the system provides the most accurate and relevant answers for the user querying the data. Findings – The case demonstrated the potential of using linked data principles and the semantic search system to facilitate communication between different stakeholders. With the implementation of the indicators, the system ensured that the most suitable answer was always provided at the time of the query. Even if the information queried did not yet exist in the graph, the system attempted to find the best available match. Additionally, the system proved effective in knowledge discovery by suggesting contextually relevant information that was not directly queried. Research limitations/implications – This research shows promise in improving information management in construction projects, but some limitations remain. The indicators used in our assessment framework are designed to support information retrieval, but certain parameters, such as the exponents, are arbitrarily set and require optimization. Additionally, while only three indicators are currently being tested, incorporating other relevant indicators could further enhance the system's efficiency and effectiveness. Practical implications – Our approach offers an efficient semantic search system specifically tailored for the construction field; however, the query assessment framework (indicators) we created can be applied to any semantic search, as we attempt to provide a method for assessing the relevance and quality of semantic searches. Social implications – By improving the assessment of semantic search results, our approach enhances information accessibility and decision-making across various domains. This contributes to more effective knowledge sharing, reducing miscommunication and inefficiencies in collaborative environments. Originality/value – A real case study implementing linked data in renovation project and the development of a novel framework for query evaluation consisting of three main metrics. While each metric has been individually discussed in the literature, the relationship between actor, information and process has not been described or quantified as indicators to ensure the semantic search quality.