This study explores the history of multi-ethnic interaction and the path of cultural integration from the perspective of national unity education, combining association rule mining and questionnaire survey. Based on Apriori algorithm for association term analysis, text classification is carried out using plain Bayesian network. Based on the improved Apriori text mining model, combined with the field questionnaire survey data and the historical ethnic interaction event database, the strong correlation rules such as “traditional festival participation → crossethnic dinner” (confidence 83.1%) and “keeping promises → employment mutual assistance” (78.3%) were excavated, and it was found that the average confidence of food culture rules (0.79) was significantly higher than that of religious ritual rules (0.64), revealing the important role of material and cultural exchange in ethnic integration. Based on the comparative experimental results, the detection accuracy of the improved Apriori algorithm is verified. The data of 790 questionnaires show that the awareness of “religious belief” and “customs” of various ethnic groups constitutes the most significant difference dimension, and the average awareness of the difference in “customs and habits” reaches 60.3%. Ethnic interaction attitudes show a high degree of consistency, with more than 95% of the respondents taking honesty and reliability as the main preferred criteria for choosing friends. Accordingly, this paper proposes a third-order ethnic interaction history and cultural integration path, which provides a methodological support with both spatio-temporal precision and cultural depth for multi-ethnic interaction in the new era.
