首頁 > 網路資源 > 大同大學數位論文系統

Title page for etd-0904118-145620


URN etd-0904118-145620 Statistics This thesis had been viewed 79 times. Download 0 times.
Author Ming-Si Zou
Author's Email Address No Public.
Department Information Management
Year 2017 Semester 2
Degree Master Type of Document Master's Thesis
Language zh-TW.Big5 Chinese Page Count 47
Title Financial Customer Classification with Deep Learning using TensorFlow
Keyword
  • Direct marketing
  • Deep learning
  • TensorFlow
  • Customer classification
  • Customer classification
  • TensorFlow
  • Deep learning
  • Direct marketing
  • Abstract In financial industry, it needs to establish a model for target customers classification in order to predict whether the customers willing to buy a product. Building a customer classification model requires an efficient way to handle large amounts of various data and multiple iterative computation. This study uses deep learning to construct a customer classification model. The purpose is to observe whether deep learning can solve financial customers classification problems effectively. In experimental phase, we use TensorFlow framework which can arrange computational graphs for designing the deep learning processes and also complete a large-scale classification operations with less time and cost. This study uses a financial customer dataset to evaluate the classification performance of different deep neural network architecture. From the experimental results, the best neural network is using three hidden layer and 60 neurons inside the hidden layers. In addition, we adjust the dropout parameter at 0.6 to prevent over-fitting.
    Advisor Committee
  • Yucheng Kao - advisor
  • none - co-chair
  • none - co-chair
  • Files indicate accessible at a year
    Date of Defense 2018-07-04 Date of Submission 2018-09-04


    Browse | Search All Available ETDs