eva‎luation of the Logistics Performance Index (LPI): An Analysis of Hybrid Methods Using Text Mining (LDA) an‎d Structural Equation Modeling (SEM)

2/19/2026 12:00:00 AM

In this thesis, the Logistics Performance Index (LPI) is analyzed as one of the key indicators for eva‎luating the Logistics performance of countries. The research approach utilizes the Latent Dirichlet Allocation (LDA) model for text analysis an‎d identifying the factors influencing Logistics performance. The data used is extracted from reliable international sources such as the International Monetary Fund, the World Trade Organization, an‎d Sustainable Development Reports. Subsequently, using Structural Equation Modeling (SEM), the relationships between latent variables an‎d observed variables were analyzed an‎d modeled. The results indicated that economic factors, Logistics an‎d transportation infrastructure, policy-making, an‎d environmental sustainability are among the most influential factors on the Logistics performance index. However, weaknesses in environmental factors highlight the need to revise the criteria an‎d improve environmental approaches. The results of text mining led to the identification of ten main topics related to the Logistics Performance Index (LPI). These topics represent various aspects of trade, infrastructure, policy-making, an‎d economic an‎d environmental sustainability. However, in the Structural Equation Modeling (SEM) phase, due to the importance of model integration an‎d simplification, as well as avoiding overlap between variables, the ten initial topics were consolidated into eight latent variables. In this study, Logistics was analyzed as the primary variable, an‎d factors such as "Logistics an‎d transportation infrastructure," "operational performance," an‎d "policy-making an‎d governance" were identified as influential factors. The results showed that "Logistics an‎d transportation infrastructure" had the greatest direct impact on "national Logistics efficiency." Moreover, international transport indicators an‎d port capacity play a key role in this area. Additionally, the quality of cargo tracking, scheduling accuracy, an‎d customs efficiency can enhance Logistics performance an‎d reduce trade barriers, with effective customs policies an‎d greater transparency in regulations having a strong positive impact on this index. This research, by providing a novel framework, offers tools for a deeper an‎d more precise analysis of Logistics performance for policymakers an‎d researchers, emphasizing the importance of comprehensive approaches in eva‎luating Logistics performance. It also provides recommendations for improving policies, developing infrastructure, an‎d strengthening Logistics sustainability at the national level.

شاخص عملكرد لجستيك , متن كاوي , تخصيص ديريكله نهفته , مدل‌سازي معادلات ساختاري , لجستيك پايدار

Logistics Performance Index (LPI) , Text Mining , Latent Dirichlet Allocation (LDA) , Structural Equation Modeling (SEM) , Sustainable Logistics

Pantea Saboktakin

Morteza Bagheri