Recommending Judgment Targets for Rating Provision

Existing rating and reviewing schemes typically come in the in the flavor of a single rating and/or a textual review. While a single judgment evaluating the overall quality of a product is of limited significance, textual customer reviews typically deliver more informative feedback at the attribute level. However, reading and comparing them to extract relevant information is time-consuming and mentally-demanding. Moreover, they are often biased and selective in the product features they consider and thus are less helpful for making informed buying decisions. We therefore propose a rating elicitation scheme that supports consumers in providing meaningful and machine-comprehensible, i.e automatically processable, responses in terms of multi-criteria ratings judging several features of a purchased product or used service. This is achieved by recommending suitable judgment targets and thereby accounting for a customer’s willingness to provide ratings. Our evaluation results show that the proposed procedure effectively adjusts to a consumer’s personal judgment preferences and thus provides helpful support for the elicitation of meaningful multi-criteria feedback.