Abstract: Firstly, the “platform user” two-way interactive comprehensive performance evaluation index system was constructed from three aspects of platform resources, user experience and service efficiency, and the platform performance evaluation ontology base was constructed based on this. Secondly, for the important data source user comments in the index system, the ontology base was used as the abstract knowledge to map the word segmentation feature items in the user comments to the ontology base for semantic similarity calculation, so as to realize the semantic analysis and understanding of the user comments, and carry out numerical extraction or hierarchical classification of the indicators involved in the user comments, so as to provide a rich and accurate user feedback data set for the platform performance evaluation. Finally, the weight of each index of performance evaluation was determined by using the analytic hierarchy process, and the performance value of the agricultural comprehensive information service platform was obtained by using the weighted calculation method to quantitatively evaluate the agricultural information service effectiveness of each provincial platform. The results showed that platform resources and user experience had a great impact on the performance of the agricultural comprehensive information service platform. The performance of agricultural comprehensive information service platforms in Jiangsu Province, Henan Province and Yunnan Province were 3.17, 3.25 and 2.86 respectively. Experience has proved that the evaluation method of this study had practical value.

Key words: agricultural comprehensive information service platform, performance evaluation, ontology modeling, evaluating indicator