Unlocking Second-hand Sailboat Prices: A Data-Driven Approach

Yinlu Xia

DOI: https://doi.org/10.54097/hbem.v20i.14683

2023-11-30

Abstract:With the rapid advancements in science and technology, the manufacturing capacity of sailboats has significantly improved, leading to an expansion in the second-hand sailboat market. However, the absence of a comprehensive pricing mechanism for second-hand sailboats remains a challenge. This study aims to address this pricing issue for sailing vessels by establishing a robust mathematical model. The initial step in this investigation involved preprocessing all provided data. As a preliminary example, a one-way analysis of variance revealed a substantial influence of the sailboat brand on pricing. Consequently, a K-means clustering analysis was conducted, resulting in the categorization of sailboats into 20 distinct classes. To streamline subsequent analyses, the most numerous variants within each class were selected as representative samples. Further, relevant data on length overallLOA), length at waterlineLWL), and other numerical indicators were extracted. Similarly, for effective quantification of all textual data, this research opted to supplement geographic information such as northern latitudes and GDP. Subsequently, a multiple linear regression analysis was executed on the entire dataset, yielding the final pricing formula as the model. This methodology was applied to catamarans as well. Additionally, this paper explores the potential impact of geographical factors on pricing by examining various regions. Parameters including coastline length, GDP, wind and wave conditions, weather, and sea transport were selected for quantification. These factors were further assessed through principal component analysis to derive composite factors, whose weights were determined using the TOPSIS entropy weight method. These weights signify the degree of influence on pricing. To investigate regional effects, the contribution of geographical factors to sailboat pricing was assessed. Ultimately, the variables deemed to have significant practical implications were identified.