Identifying User Profile by Incorporating Self-Attention Mechanism Based on CSDN Data Set

Five tables and four figures of this paper. Table 1 shows statistics of the evaluation data set. Table 2 presents a sample of CSDN data set. Table 3 is a comparison on Task 1 with different aspects. Table 4 is a comparison of different aspects on Task 2. Table 5 shows performance of UIR-SIST system in SMP CUP 2017. Figure 1 shows a system architecture. Figure 2 is a framework of CNNs model based on weighted-blog-embeddings in Task 2. Figure 3 presents a framework of the stacking model in Task 3. Figure 4 shows an example of daily statistics of user behaviors. Note: “Add” refers to “add favoriates”, and “send” refers to “send private messages”.

本文共包含五张表格与四幅图表。表1展示了评估数据集的统计信息。表2展示了CSDN数据集的样本。表3对Task 1的不同方面进行了对比分析。表4对Task 2的不同方面进行了比较。表5呈现了UIR-SIST系统在2017年SMP CUP竞赛中的性能表现。图1展示了系统架构。图2展示了基于加权博客嵌入的CNNs模型框架，应用于Task 2。图3呈现了Task 3中堆叠模型的框架。图4展示了用户行为日统计数据的示例。注释：'Add'指'添加收藏夹'，'send'指'发送私信'。