新闻分类

赛题数据为新闻文本，并按照字符级别进行匿名处理。整合划分出14个候选分类类别：财经、彩票、房产、股票、家居、教育、科技、社会、时尚、时政、体育、星座、游戏、娱乐的文本数据。 赛题数据由以下几个部分构成：训练集20w条样本，测试集A包括5w条样本The competition dataset consists of news texts anonymized at the character level. These texts are categorized into 14 candidate classes: Finance, Lottery, Real Estate, Stocks, Home & Household, Education, Technology, Society, Fashion, Current Politics, Sports, Horoscope, Games, and Entertainment. The competition dataset comprises two components: a training set with 200,000 samples, and Test Set A containing 50,000 samples.

搜集汇总

背景与挑战

背景概述

该数据集包含经过字符匿名处理的新闻文本，覆盖财经、彩票、房产等14个分类类别。数据由20万条训练样本和5万条测试样本（测试集A）构成。