The traditional learning resources system gathered a mass of heterogeneous data. It is deficient in intelligence and personalization. Its development is complicated and long period. Learners acquire desired resources only through tedious frequent search. In view of shortcomings found in traditional learning resources system, this paper applied data mining technologies to the research of intelligent push while using Ruby on Rails' full-stack MVC architecture for the system development. The system framework and algorithm is designed in this paper. Studies show that the intelligent push technologies and agile development methods can effectively improve the efficiency of learning resources utilization and cut down the development cycle.

In this work, the interdiffusion coefficient and atomic mobility for fcc Ag-Cu-Mg phase at 1073 K have been investigated by combining diffusion couple experiments and calculations. Based on the experimental composition profiles, the diffusion coefficients at intersection points in diffusion paths were calculated by means of Matano-Kirkaldy method. Using the thermodynamic descriptions available in literature, the atomic mobilities for fcc Ag-Cu-Mg phase were automatically optimized by numerical inverse method incorporated in the CALTPP (CALculation of ThermoPhysical Properties) program. Moreover, the obtained atomic mobilities were verified to be reliable through good agreements between the model-simulated composition profiles and the measured ones. In addition, three-dimensional surfaces for the interdiffusion coefficient, activation energy, and frequency factor were presented. The presently obtained atomic mobilities can be incorporated into the diffusion database for Ag-based alloys, which can contribute to microstructure simulation and materials design.