SGComp data sets

Training data set for an expert finding system (SGComp_Train.csv), and Benchmarking data set for expert finding systems with adversarial data (SGComp_Adversarial.csv). Both data sets have the same field structure as relation triples (head: experts' full name, relation: relationship with key phrases, list of tail: []) in comma-separated values (.CSV)本数据集包含两类资源：一是面向专家检索系统的训练数据集（SGComp_Train.csv），二是面向带对抗数据的专家检索系统的基准测试数据集（SGComp_Adversarial.csv）。 两类数据集均采用逗号分隔值（CSV）格式，其字段结构遵循关系三元组范式：头实体（head）为专家全名，关系（relation）为与关键短语的关联关系，尾实体列表（tail: []）为列表格式。