Abstract

A comprehensive profile database of industrial wastewater in Taiwan has been developed. The dataset consists of 244 wastewater samples containing 52 metals and metalloid elements. A Naïve Bayes classifier was applied to establish robust composite profiles for each industrial category. Similarities in the composite profiles were examined using a hierarchical cluster analysis, and most composite profiles were distinct. Finally, 19 composite profiles of the industrial categories were subsequently retrieved and utilized to identify the marker species for each category. During the examination of marker species through statistical methods, three less noticeable characteristic species, Mo, Sb, and Co, were examined further: Molybdenum was confirmed as a marker species for the petrochemical industry, while antimony was identified as a marker for man-made fibers and finishing of textiles. Cobalt was also determined as a marker for nickel-based electroplating products. We compared the regulations of other countries with those of Taiwan in order to offer scientific advice regarding these three species. This study serves as an exemplary application of a profile database, showing how it can be used to analyze less noticeable marker species among the 52 elements for different industries. Moreover, the study results suggest that composite profiles generated from various industries hold potential as tools for future river pollution prevention. Overall, the continuous expansion of the database is highly recommended to increase its practicality and relevance for future research and regulatory efforts.