Cross-border e-commerce commodity risk assessment using text mining and fuzzy rule-based reasoning

Abstract

The potentially conflicting regulations of different countries can cause significant risk for cross-border e-commerce (CBEC), since the multi-varieties, small-batch transaction mode of CBEC prohibit the customs from carrying out comprehensive inspection on commodity quality. In this paper, a risk assessment method utilizing text mining and fuzzy rule-based reasoning is proposed, which is able to assess CBEC commodity risk quantitatively and semi-automatically. To do this, first the commodity safety regulations are transformed into risk rules considering a fuzzy representation of risk consequence. Then the features of CBEC commodities are extracted from webpages and the commodities are classified to be assessed with corresponding risk factors. For the inherent risk factors representing violations of safety regulations, the risk consequence is inferred using fuzzy rules. For the historical risk factors representing how often a kind of commodities have violated safety regulations, the risk probability and consequence are assessed by analyzing the import inspection record. At last, the overall risk of a CBEC commodity is obtained by aggregating the output of related risk factor. Experimental results show that the proposed method enhances CBEC commodity risk assessment by improving its efficiency and accuracy.