Dataset Preview Go to dataset viewer
utterance (string)intent (string)category (string)tags (string)
"would it be possible to cancel the order I made?"
"cancel_order"
"ORDER"
"BIP"
"cancelling order"
"cancel_order"
"ORDER"
"BK"
"I need assistance canceling the last order I have made"
"cancel_order"
"ORDER"
"B"
"problem with canceling the order I made"
"cancel_order"
"ORDER"
"B"
"I don't know how to cancel the order I made"
"cancel_order"
"ORDER"
"B"
"can you help me cancel the order I made?"
"cancel_order"
"ORDER"
"BI"
"I would like to know about order cancellations"
"cancel_order"
"ORDER"
"BMP"
"could you help me cancelling an order?"
"cancel_order"
"ORDER"
"BIP"
"I don't know how to cancel an order I made"
"cancel_order"
"ORDER"
"B"
"help me cancelling my last order"
"cancel_order"
"ORDER"
"B"
"I do not know how to cancel the last order I made"
"cancel_order"
"ORDER"
"BE"
"I need assistance with canceling an order I made"
"cancel_order"
"ORDER"
"B"
"information about canceling an order"
"cancel_order"
"ORDER"
"B"
"I would like to cancel an order"
"cancel_order"
"ORDER"
"BP"
"I need assistance with cancelling my orders"
"cancel_order"
"ORDER"
"BM"
"I need assistance with canceling the order I have made"
"cancel_order"
"ORDER"
"B"
"assistance canceling the order I have made"
"cancel_order"
"ORDER"
"B"
"question about cancelling the last order"
"cancel_order"
"ORDER"
"B"
"help me to cancel my last order"
"cancel_order"
"ORDER"
"B"
"problem with cancelling orders"
"cancel_order"
"ORDER"
"BM"
"problems with cancelling an order"
"cancel_order"
"ORDER"
"BM"
"assistance with cancelling the order I made"
"cancel_order"
"ORDER"
"B"
"I need help with cancelling an order I have made"
"cancel_order"
"ORDER"
"B"
"I have problems with cancelling my order"
"cancel_order"
"ORDER"
"BM"
"I do not want to pay for the order I made"
"cancel_order"
"ORDER"
"BE"
"I have a problem with cancelling orders"
"cancel_order"
"ORDER"
"BM"
"problems with canceling orders"
"cancel_order"
"ORDER"
"BM"
"where can I get information about canceling orders?"
"cancel_order"
"ORDER"
"BIM"
"I want assistance cancelling an order I have made"
"cancel_order"
"ORDER"
"B"
"cancel order"
"cancel_order"
"ORDER"
"BK"
"find information about order cancelations"
"cancel_order"
"ORDER"
"BM"
"can you give me information about cancelling orders?"
"cancel_order"
"ORDER"
"BIM"
"I have a question about cancelling my last order"
"cancel_order"
"ORDER"
"B"
"I don't know how to cancel my last order"
"cancel_order"
"ORDER"
"B"
"I have problems with canceling my last order"
"cancel_order"
"ORDER"
"BM"
"problem with cancelling an order I have made"
"cancel_order"
"ORDER"
"B"
"I need help with canceling the last order I made"
"cancel_order"
"ORDER"
"B"
"how can I find information about cancelling orders?"
"cancel_order"
"ORDER"
"BIM"
"I want to know more about order cancelations"
"cancel_order"
"ORDER"
"BM"
"I want help to cancel an order I have made"
"cancel_order"
"ORDER"
"B"
"I want assistance canceling an order"
"cancel_order"
"ORDER"
"B"
"can you help me to cancel the last order I have made?"
"cancel_order"
"ORDER"
"BI"
"give me information about cancelling orders"
"cancel_order"
"ORDER"
"BM"
"help cancelling my order"
"cancel_order"
"ORDER"
"B"
"I have a question about canceling my order"
"cancel_order"
"ORDER"
"B"
"help me canceling my last order"
"cancel_order"
"ORDER"
"B"
"would it be possible to cancel my order?"
"cancel_order"
"ORDER"
"BIP"
"where could I get information about cancelling an order?"
"cancel_order"
"ORDER"
"BIP"
"could you help me cancelling an order I made?"
"cancel_order"
"ORDER"
"BIP"
"where can I get information about canceling an order?"
"cancel_order"
"ORDER"
"BI"
"I have to cancel my last order"
"cancel_order"
"ORDER"
"B"
"question about canceling my order"
"cancel_order"
"ORDER"
"B"
"assistance cancelling my order"
"cancel_order"
"ORDER"
"B"
"I have problems with canceling the order I have made"
"cancel_order"
"ORDER"
"BM"
"help canceling the last order"
"cancel_order"
"ORDER"
"B"
"help canceling an order"
"cancel_order"
"ORDER"
"B"
"I try to cancel the order I made"
"cancel_order"
"ORDER"
"B"
"I need to cancel my order"
"cancel_order"
"ORDER"
"B"
"could you help me cancelling the last order I made?"
"cancel_order"
"ORDER"
"BIP"
"where could I cancel the order I made?"
"cancel_order"
"ORDER"
"BIP"
"I don't know how I could cancel an order I have made"
"cancel_order"
"ORDER"
"BP"
"I have problems with cancelling the last order I made"
"cancel_order"
"ORDER"
"BM"
"help me to cancel an order I have made"
"cancel_order"
"ORDER"
"B"
"I am trying to find information about order cancelations"
"cancel_order"
"ORDER"
"BEM"
"where can I cancel my order?"
"cancel_order"
"ORDER"
"BI"
"I need assistance canceling the order I made"
"cancel_order"
"ORDER"
"B"
"would it be possible to cancel the last order I have made?"
"cancel_order"
"ORDER"
"BIP"
"I have a question about canceling the last order"
"cancel_order"
"ORDER"
"B"
"I need help to cancel the order I made"
"cancel_order"
"ORDER"
"B"
"I do not know how I could cancel an order"
"cancel_order"
"ORDER"
"BEP"
"cancel an order I have made"
"cancel_order"
"ORDER"
"B"
"help with cancelling the order I have made"
"cancel_order"
"ORDER"
"B"
"I can't pay for the order I have made"
"cancel_order"
"ORDER"
"B"
"I can't pay for the order"
"cancel_order"
"ORDER"
"B"
"I need help cancelling an order I made"
"cancel_order"
"ORDER"
"B"
"would you give me information about canceling orders?"
"cancel_order"
"ORDER"
"BIMP"
"help me canceling the last order I have made"
"cancel_order"
"ORDER"
"B"
"I am looking for information about order cancelations"
"cancel_order"
"ORDER"
"BEM"
"I don't know how I can cancel the last order I have made"
"cancel_order"
"ORDER"
"B"
"how could I cancel the last order I have made?"
"cancel_order"
"ORDER"
"BIP"
"is it possible to cancel my order?"
"cancel_order"
"ORDER"
"BI"
"help cancelling the last order I have made"
"cancel_order"
"ORDER"
"B"
"I have a question about canceling orders"
"cancel_order"
"ORDER"
"BM"
"how to cancel the last order I have made?"
"cancel_order"
"ORDER"
"BI"
"I don't want an order I have made"
"cancel_order"
"ORDER"
"B"
"I have a question about cancelling my order"
"cancel_order"
"ORDER"
"B"
"how could I cancel an order I made?"
"cancel_order"
"ORDER"
"BIP"
"I want help cancelling an order I have made"
"cancel_order"
"ORDER"
"B"
"I have problems with cancelling the order I have made"
"cancel_order"
"ORDER"
"BM"
"how do I cancel my order?"
"cancel_order"
"ORDER"
"BI"
"could you help me to cancel my last order?"
"cancel_order"
"ORDER"
"BIP"
"can you help me cancel my last order?"
"cancel_order"
"ORDER"
"BI"
"I do not want to pay for an order I made"
"cancel_order"
"ORDER"
"BE"
"assistance cancelling the last order"
"cancel_order"
"ORDER"
"B"
"assistance cancelling an order I made"
"cancel_order"
"ORDER"
"B"
"help with canceling the order I have made"
"cancel_order"
"ORDER"
"B"
"would you give me information about order cancelations?"
"cancel_order"
"ORDER"
"BIMP"
"where could I cancel the last order?"
"cancel_order"
"ORDER"
"BIP"
"I need help to cancel the last order I have made"
"cancel_order"
"ORDER"
"B"
"will you give me information about canceling an order?"
"cancel_order"
"ORDER"
"BIP"
End of preview (truncated to 100 rows)
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/datasets-cards)

Bitext - Customer Service Tagged Training Dataset for Intent Detection

Overview

The dataset can be used to train intent recognition models on Natural Language Understanding (NLU) platforms: LUIS, Dialogflow, Lex, RASA and more. The dataset covers the "Customer Service" domain and includes:

  • 11 categories or intent groups
  • 27 intents assigned to one of the 11 categories
  • 8,175 utterances assigned to the 27 intents

Additionally, each utterance is enriched with tags that indicate the type of language variation that the utterance expresses. Examples include:

  • The tag “COLLOQUIAL” indicates that the utterance contains informal expressions: “can u close my account”
  • The tag “INTERROGATIVE” indicates that the utterance is a question: “how do I open an account”
  • The tag “OFFENSIVE” indicates that the utterance contains offensive expressions: “open my f****** account”

There are a total of 11 tags. See below for a full list of tags, categories and intents.

The purpose of these tags is to customize the dataset so the trained bot can easily adapt to different user language profiles. A bot that sells sneakers and targets a younger population should be proficient in colloquial language; while a classical retail banking bot should be able to handle more formal or polite language.

These intents have been selected from Bitext's collection of 20 domain-specific datasets (banking, retail, utilities...), covering the intents that are common across all 20 domains. For a full list of domains see https://www.bitext.com/chatbot-verticals/.

Utterances and Linguistic Tags

The dataset contains 8,175 training utterances, with between 290 and 324 utterances per intent.

The dataset has been split into training (80%), validation (10%) and testing (10%) sets, preserving the distribution of intents and linguistic phenomena.

The dataset also reflects commonly occurring linguistic phenomena of real-life chatbots, such as: spelling mistakes, run-on words, punctuation errors…

Each entry in the dataset contains the following four fields:

  • utterance: a user utterance from the Customer Service domain
  • intent: the intent corresponding to the user utterance
  • category: the high-level semantic category for the intent
  • tags: different tags that reflect the types of language variations expressed in the utterance

The dataset contains tags that reflect different language phenomena like colloquial or offensive language. So if an utterance for intent “cancel_order” contains the “COLLOQUIAL” tag, the utterance will express an informal language variation like: “can u cancel my order”

Each utterance is enriched with one or more of these tags:

  • Register tags: colloquial language, polite language…
    • Q - Colloquial variation
    • P - Politeness variation
  • Content tags: offensive language, keyword language…
    • W - Offensive language
    • K - Keyword language
  • Linguistic tags: syntactic and morphological tags (interrogative sentence, coordinated sentence…)
    • B - Basic syntactic structure
    • C - Coordinated syntactic structure
    • I - Interrogative structure
    • M - Morphological variation (plurals, tenses…)
    • L - Lexical variation (synonyms)
    • E - Expanded abbreviations (I'm -> I am, I'd -> I would…)
  • Real-life errors: spelling errors, punctuation errors…
    • Z - Noise phenomena like spelling or punctuation errors

These tags indicate the type of language variation that the utterance expresses. When associated to each utterance, they allow Conversational Designers to customize training datasets to different user profiles with different uses of language. Through these tags, many different datasets can be created to make the resulting assistant more accurate and robust. A bot that sells sneakers should be mainly targeted to younger population that use a more colloquial language; while a classical retail banking bot should be able to handle more formal or polite language.

Categories and Intents

The categories and intents covered by the dataset are:

  • ACCOUNT: create_account, delete_account, edit_account, recover_password, registration_problems, switch_account
  • CANCELLATION_FEE: check_cancellation_fee
  • CONTACT: contact_customer_service, contact_human_agent
  • DELIVERY: delivery_options, delivery_period
  • FEEDBACK: complaint, review
  • INVOICE: check_invoice, get_invoice
  • NEWSLETTER: newsletter_subscription,
  • ORDER: cancel_order, change_order, place_order, track_order
  • PAYMENT: check_payment_methods, payment_issue
  • REFUND: check_refund_policy, get_refund, track_refund
  • SHIPPING_ADDRESS: change_shipping_address, set_up_shipping_address

(c) Bitext Innovations, 2022

Edit dataset card
Evaluate models HF Leaderboard