地球资源数据云——数据资源详情
南方公园对话
该数据集《South Park Dialogue》主要用于监督学习任务，数据形态以文本为主，应用场景偏向交通/汽车。 题目说明：More than 70,000 lines of dialogue by season, episode, and character 任务类型：文本监督学习。 建议流程：先做文本清洗与分词，再比较 TF - IDF+线性模型 与 预训练语言模型。 评估建议：使用分层切分或交叉验证，优先关注 F1、Recall、AUC 等分类指标。 可用文件：All - seasons.csv。 South Park cartoon lines +70k lines, annotated with season, episode and speaker It is interesting to practice NLP with ML techniques in order to guess who is speaking. Later there will be file with pre - proccesed data to train
摘要概览
该数据集《South Park Dialogue》主要用于监督学习任务，数据形态以文本为主，应用场景偏向交通/汽车。 题目说明：More than 70,000 lines of dialogue by season, episode, and character
任务类型：文本监督学习。
建议流程：先做文本清洗与分词，再比较 TF - IDF+线性模型 与 预训练语言模型。
评估建议：使用分层切分或交叉验证，优先关注 F1、Recall、AUC 等分类指标。
可用文件：All - seasons.csv。
South Park cartoon lines
+70k lines, annotated with season, episode and speaker
It is interesting to practice NLP with ML techniques in order to guess who is speaking. Later there will be file with pre - proccesed data to train