id
stringlengths
19
19
target
sequence
label
sequence
user_id_list
sequence
1494101762587836419
[ 3, 1, 3 ]
[ 4, 3, 4 ]
[ 65, 90, 51 ]
1494127310340976648
[ 1, 1, 1 ]
[ 2, 2, 2 ]
[ 65, 90, 51 ]
1494128253279567872
[ 3, 1, 1 ]
[ 1, 1, 1 ]
[ 65, 90, 51 ]
1494131907076190209
[ 1, 3, 1 ]
[ 1, 4, 1 ]
[ 65, 90, 51 ]
1494150116118515715
[ 1, 1, 3 ]
[ 3, 3, 4 ]
[ 65, 90, 51 ]
1494156454793969666
[ 1, 1, 3 ]
[ 4, 2, 2 ]
[ 65, 90, 51 ]
1494162318527463433
[ 3, 3, 3 ]
[ 3, 4, 2 ]
[ 65, 90, 51 ]
1494162638825467905
[ 3, 1, 3 ]
[ 3, 2, 3 ]
[ 65, 90, 51 ]
1494167418121637889
[ 2, 1, 1 ]
[ 3, 2, 2 ]
[ 65, 90, 51 ]
1494176257046908933
[ 1, 1, 3 ]
[ 4, 2, 4 ]
[ 65, 90, 51 ]
1494191889574084608
[ 1, 1, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494239227416297472
[ 1, 1, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494241642215133189
[ 1, 1, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494243758572933120
[ 1, 1, 1 ]
[ 3, 2, 2 ]
[ 65, 90, 51 ]
1494249749272092682
[ 2, 2, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494275699456106496
[ 1, 1, 1 ]
[ 1, 1, 2 ]
[ 65, 90, 51 ]
1494275950032273409
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494298478746701825
[ 3, 1, 1 ]
[ 3, 1, 2 ]
[ 65, 90, 51 ]
1494300906149482498
[ 1, 3, 1 ]
[ 3, 4, 3 ]
[ 65, 90, 51 ]
1494307282171363332
[ 1, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494313039692460040
[ 1, 3, 1 ]
[ 3, 4, 2 ]
[ 65, 90, 51 ]
1494316978135535618
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494318450826948611
[ 1, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494319789212581892
[ 1, 1, 1 ]
[ 2, 3, 2 ]
[ 65, 90, 51 ]
1494320352628985857
[ 1, 0, 1 ]
[ 3, 0, 4 ]
[ 65, 90, 51 ]
1494325458036740098
[ 1, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494326477495869441
[ 1, 1, 3 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494328200947974150
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494329699920904196
[ 1, 1, 1 ]
[ 2, 3, 3 ]
[ 65, 90, 51 ]
1494335278450094083
[ 3, 1, 1 ]
[ 4, 3, 2 ]
[ 65, 90, 51 ]
1494377364532318217
[ 0, 1, 3 ]
[ 0, 3, 3 ]
[ 65, 90, 51 ]
1494383127304687621
[ 3, 3, 1 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494386642672832512
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494393896419291137
[ 1, 1, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494398845253722112
[ 1, 1, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494409179645509632
[ 0, 0, 0 ]
[ 0, 0, 0 ]
[ 65, 90, 51 ]
1494428047487156227
[ 1, 3, 1 ]
[ 2, 4, 2 ]
[ 65, 90, 51 ]
1494432035640786958
[ 1, 3, 1 ]
[ 2, 4, 2 ]
[ 65, 90, 51 ]
1494435703207645185
[ 1, 3, 1 ]
[ 3, 4, 3 ]
[ 65, 90, 51 ]
1494447675885821952
[ 1, 3, 1 ]
[ 1, 4, 1 ]
[ 65, 90, 51 ]
1494452454385393664
[ 1, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494452673432547332
[ 1, 1, 1 ]
[ 2, 3, 2 ]
[ 65, 90, 51 ]
1494455157735981058
[ 1, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494455982725816322
[ 1, 3, 1 ]
[ 4, 4, 2 ]
[ 65, 90, 51 ]
1494456238733955072
[ 3, 1, 1 ]
[ 1, 2, 2 ]
[ 65, 90, 51 ]
1494459153498402820
[ 1, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494459193151332353
[ 2, 3, 1 ]
[ 1, 2, 2 ]
[ 65, 90, 51 ]
1494459486521937921
[ 1, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494460579435917312
[ 1, 1, 2 ]
[ 2, 3, 1 ]
[ 65, 90, 51 ]
1494472083681603590
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494531282663256066
[ 1, 1, 3 ]
[ 4, 3, 4 ]
[ 65, 90, 51 ]
1494546353913200645
[ 1, 1, 1 ]
[ 2, 3, 2 ]
[ 65, 90, 51 ]
1494579949529079813
[ 1, 3, 3 ]
[ 4, 3, 4 ]
[ 65, 90, 51 ]
1494583604785352706
[ 2, 1, 2 ]
[ 3, 2, 2 ]
[ 65, 90, 51 ]
1494592954643611651
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494595781516722177
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494620913501085700
[ 1, 1, 1 ]
[ 1, 3, 2 ]
[ 65, 90, 51 ]
1494629786878025730
[ 2, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494633554529374208
[ 1, 1, 1 ]
[ 4, 3, 2 ]
[ 65, 90, 51 ]
1494638550893330432
[ 1, 0, 1 ]
[ 2, 0, 2 ]
[ 65, 90, 51 ]
1494643413265702913
[ 1, 1, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494648386892759040
[ 2, 3, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494650742451638272
[ 1, 1, 1 ]
[ 3, 2, 2 ]
[ 65, 90, 51 ]
1494662840162336769
[ 3, 1, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494669078279364609
[ 0, 1, 3 ]
[ 0, 3, 4 ]
[ 65, 90, 51 ]
1494675239699365888
[ 1, 1, 1 ]
[ 1, 3, 2 ]
[ 65, 90, 51 ]
1494686530832859137
[ 1, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494688036298567684
[ 1, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494695124450025472
[ 1, 3, 1 ]
[ 4, 4, 2 ]
[ 65, 90, 51 ]
1494695964363296774
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494702037744820225
[ 1, 3, 1 ]
[ 3, 4, 3 ]
[ 65, 90, 51 ]
1494703453410828290
[ 1, 1, 3 ]
[ 4, 3, 4 ]
[ 65, 90, 51 ]
1494703838481186818
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494706632101888001
[ 1, 3, 1 ]
[ 1, 1, 2 ]
[ 65, 90, 51 ]
1494738482882748420
[ 3, 1, 1 ]
[ 4, 1, 2 ]
[ 65, 90, 51 ]
1494744282615353348
[ 1, 1, 1 ]
[ 3, 2, 3 ]
[ 65, 90, 51 ]
1494765432263168003
[ 1, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494780533293064192
[ 3, 1, 3 ]
[ 3, 3, 4 ]
[ 65, 90, 51 ]
1494782763022839808
[ 1, 1, 1 ]
[ 2, 2, 2 ]
[ 65, 90, 51 ]
1494784933188947968
[ 1, 1, 3 ]
[ 2, 2, 2 ]
[ 65, 90, 51 ]
1494800884982710273
[ 1, 1, 1 ]
[ 2, 2, 2 ]
[ 65, 90, 51 ]
1494804141352226817
[ 0, 3, 3 ]
[ 0, 4, 4 ]
[ 65, 90, 51 ]
1494810938146394114
[ 1, 1, 1 ]
[ 1, 2, 1 ]
[ 65, 90, 51 ]
1494811814634594304
[ 1, 1, 3 ]
[ 3, 3, 4 ]
[ 65, 90, 51 ]
1494813598346268672
[ 1, 1, 1 ]
[ 3, 3, 3 ]
[ 65, 90, 51 ]
1494814044385349632
[ 3, 1, 1 ]
[ 3, 1, 3 ]
[ 65, 90, 51 ]
1494816934558601219
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494817831988387841
[ 1, 1, 1 ]
[ 2, 2, 2 ]
[ 65, 90, 51 ]
1494820802075959296
[ 1, 1, 1 ]
[ 3, 2, 2 ]
[ 65, 90, 51 ]
1494822848141340672
[ 1, 3, 1 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494857971255312384
[ 1, 3, 1 ]
[ 1, 4, 2 ]
[ 65, 90, 51 ]
1494867779299983361
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494924448499404803
[ 3, 1, 1 ]
[ 3, 3, 2 ]
[ 65, 90, 51 ]
1494932188256604160
[ 1, 1, 1 ]
[ 3, 2, 2 ]
[ 65, 90, 51 ]
1494976700941479937
[ 1, 3, 1 ]
[ 1, 4, 1 ]
[ 65, 90, 51 ]
1494981480929382401
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]
1494986144102055938
[ 3, 3, 1 ]
[ 1, 4, 1 ]
[ 65, 90, 51 ]
1494997060369608709
[ 1, 3, 1 ]
[ 4, 4, 2 ]
[ 65, 90, 51 ]
1495007833930473478
[ 0, 3, 0 ]
[ 0, 4, 0 ]
[ 65, 90, 51 ]
1495010817582198788
[ 3, 3, 3 ]
[ 4, 4, 4 ]
[ 65, 90, 51 ]

defamation_japanese_twitter

Twitter日本語誹謗中傷検出データセット

Dataset Summary

SNSにおける誹謗中傷検出のためのデータセットです.

5,000件の日本語のツイートに,それぞれ以下で定義している誹謗中傷の対象者と内容をアノテーションしています.アノテーションは,3人のクラウドワーカーにより行われています.2022年2月15日から2022年6月30日までのツイートです. 元のツイートは含まれていないため,Twitter APIを用いてデータセットを収集してください.

中傷対象(target)と中傷内容(label)の2項目がアノテーションされています.

  • target :テキストが話題にしている対象者の分類
  • label : targetで選択された対象者に対する誹謗中傷の種類の分類

文として成立しておらず意味の取れないものはラベルC(0)としています.

target 対象
A1(1) (人種・性別・職業・思想などを共通とする)グループ (人種・性別・職業・思想などを共通とする)グループ
A2(2) 個人(著名人や知人など) 〇〇大統領,芸能人の〇〇さん,おまえ
A3(3) 対象がはっきりしないもの
C(0) 文として成立しておらず意味が取れない
label 誹謗中傷の種類 侵害されるもの
B1(1) 生命を脅かす,精神的・身体的な危害を加える 私生活の平穏 • 殺害予告などの脅迫発言
• ◯◯なんていなくなればいいのにな
B2(2) 容姿,人格などをけなしている 名誉感情 • 太っているくせにカッコいいと勘違いしている
• 田舎育ちだからファッション感覚がない
B3(3) 社会から客観的に受ける価値を低下させる 名誉権 • ◯◯さんは過去に事件を起こして逮捕されたことがある
• ◯◯さんは会社の同僚と不倫をしている
B4(4) B1-B3のどれにも当てはまらず中傷性がない
C(0) 文として成立しておらず意味が取れない

Data Fields

  • id Twitter ID
  • target: 3名のアノテータのカテゴリAの回答 values: C(0), A1(1), A2(2), A3(3)
  • label: 3名のアノテータのカテゴリBの回答 values: C(0), B1(1), B2(2), B3(3), B4(4)
  • user_id_list: 匿名化された回答者のID

Example Using Twitter API

Open In Colab

# sample code from https://github.com/twitterdev/Twitter-API-v2-sample-code/blob/main/Tweet-Lookup/get_tweets_with_bearer_token.py
import requests
import os
import json
from datasets import load_dataset

# To set your enviornment variables in your terminal run the following line:
# export 'BEARER_TOKEN'='<your_bearer_token>'
bearer_token = os.environ.get("BEARER_TOKEN")


def create_url(ids: list):
    tweet_fields = "tweet.fields=created_at"
    ids = f"ids={','.join(ids)}"
    url = "https://api.twitter.com/2/tweets?{}&{}".format(ids, tweet_fields)
    return url


def bearer_oauth(r):
    """
    Method required by bearer token authentication.
    """

    r.headers["Authorization"] = f"Bearer {bearer_token}"
    r.headers["User-Agent"] = "v2TweetLookupPython"
    return r


def connect_to_endpoint(url):
    response = requests.request("GET", url, auth=bearer_oauth)
    if response.status_code != 200:
        raise Exception(
            "Request returned an error: {} {}".format(
                response.status_code, response.text
            )
        )
    return response.json()


def get_text_data(examples):
    url = create_url(examples["id"])
    json_response = connect_to_endpoint(url)
    # print(json_response["data"])
    text_dict = {data["id"]: data["text"] for data in json_response["data"]}
    time_dict = {data["id"]: data["created_at"] for data in json_response["data"]}
    return {
        "text": [text_dict.get(id) for id in examples["id"]],
        "created_at": [time_dict.get(id) for id in examples["id"]],
    }


dataset = load_dataset("kubota/defamation-japanese-twitter")
dataset = dataset.map(get_text_data, batched=True, batch_size=100)
dataset["train"].to_pandas().head()

Contributions

Thanks to @kubotaissei for adding this dataset.

Downloads last month
53
Edit dataset card

Models trained or fine-tuned on kubota/defamation-japanese-twitter