Abstract：Exploration of usage behavioral model construction for university library electronic resources from Deep Learning Multilayer perceptron The majority of this study is to use the MLPMultilayer perceptronoperation method of deep learning to predict the readers behavior intention and the quality of use behavior generated from the electronic library service quality of the university libraryA questionnaire was built based on the quality of network electronic resource services and the theoretical structure of UTAUTCombining the public and private university students who are four-year universities and two-year masters in Taiwanthere were 1206 questionnaires issued and 1071 valid questionnaires were collectedThe predicted results of the MLP were based on the Pearson coefficient with the calculated results of the SPSS statistical methodIt was found that the calculation method of AI can accurately predict the usage behavior of 9075or moreAccording to the practiceMLP regression models can be applied to the construction of library electronic resourcessuch as the addition of electronic multimedia resourcesor the choice of books in the futureYou can use this type of deep learning model to quickly predict and make decisionsLibrary AutomationandCollection ClassificationThe purpose of this research is to use MLP to predict the intention of reader behavior and use quality generated from the quality of electronic resource services from college librariesThis means that it can be applied to the construction of the electronic resources of the librarysuch as adding electronic multimedia resources or future book choicesand then automated these processesIn additionit can be used to quickly predict and make decisions for collectionsDeep LearningThis study uses the multi-layer perceptionMLPoperation method to predict the behavior of the readers behavior and the quality of use of the quality of the service quality of the university libraryMLP is a front-oriented artificial neural networkwhich is used to generate a set of output from a set of inputsMachine LearningUse MLP for data training and testing to learn and predict the model based on this