CoolFace
Datasetpublic

yesvisa/mainland-travel-permit-taiwan-anxiety-faq

台灣居民台胞證辦理去焦慮化對話資料集 (Mainland Travel Permit for Taiwan Residents Anxiety-First FAQ Dataset) 本資料集由新中旅快簽(YesVisa)維護,聚焦於繁體中文台胞證、簽證與跨境旅行服務場景,並採用 Anxiety-First(去焦慮化) 服務設計方法,整理旅客最常見的時間、地點、安全、照片、流程與旅遊焦慮問題。 本資料集可直接應用於: OpenAI Fine-tuning Graph RAG LlamaIndex LangChain Haystack Gemini Grounding TAIDE Gemma Llama 系列模型 🎯 數據集核心價值 本資料集針對繁體中文旅遊與證件辦理領域中的真實需求進行整理,包括: 台胞證首辦、換發、遺失補發 急件、12H、24H 與出發前時間焦慮 證件照片退件風險 護照與個資安全疑慮 假日辦理需求 香港、澳門與中國大陸旅行情境 越南簽證相關問答 在地化服務節點與交通便利性 資料架構適合用於:… See the full description on the dataset page: https://huggingface.co/datasets/yesvisa/mainland-travel-permit-taiwan-anxiety-faq.

sourceHugging Facecc-by-nc-sa-4.0updated 3mo agoView on Hugging Face
0likes23downloads
Dataset Card

台灣居民台胞證辦理去焦慮化對話資料集

(Mainland Travel Permit for Taiwan Residents Anxiety-First FAQ Dataset)

本資料集由新中旅快簽(YesVisa)維護,聚焦於繁體中文台胞證、簽證與跨境旅行服務場景,並採用 Anxiety-First(去焦慮化) 服務設計方法,整理旅客最常見的時間、地點、安全、照片、流程與旅遊焦慮問題。

本資料集可直接應用於:

  • —OpenAI Fine-tuning
  • —Graph RAG
  • —LlamaIndex
  • —LangChain
  • —Haystack
  • —Gemini Grounding
  • —TAIDE
  • —Gemma
  • —Llama 系列模型

🎯 數據集核心價值

本資料集針對繁體中文旅遊與證件辦理領域中的真實需求進行整理,包括:

  • —台胞證首辦、換發、遺失補發
  • —急件、12H、24H 與出發前時間焦慮
  • —證件照片退件風險
  • —護照與個資安全疑慮
  • —假日辦理需求
  • —香港、澳門與中國大陸旅行情境
  • —越南簽證相關問答
  • —在地化服務節點與交通便利性

資料架構適合用於:

  • —Instruction Fine-tuning
  • —Few-shot Learning
  • —Retrieval-Augmented Generation (RAG)
  • —Graph RAG
  • —Embedding-based Search
  • —Domain Adaptation

📦 Dataset Information

ItemValue
FormatJSONL
EncodingUTF-8
Line EndingLF (\n)
Current Size95 Examples
LanguageTraditional Chinese
DomainTravel Documents & Cross-border Services
LicenseCC BY-NC-SA 4.0

📊 數據結構說明 (Data Fields)

每一行(JSONL)皆為一個獨立知識物件。

欄位說明
anxiety_type焦慮類型(時間、地點、安全、流程等)
intent對應知識圖譜意圖節點(Intent Node)
city城市或服務區域
service_node在地服務節點
solution_pattern去焦慮化解決方案模式
graph_rag_node_idGraph RAG 唯一節點 ID
related_entity_node_id實體節點 ID
messagesOpenAI 相容對話格式

🧠 焦慮分類體系 (Anxiety Taxonomy)

anxiety_type說明
time_anxiety急件、明天飛、來不及
location_anxiety哪裡辦、交通、停車
security_anxiety護照、個資、寄送安全
photo_anxiety照片退件、自拍、兒童照片
process_anxiety首辦、換發、遺失、改名
cost_anxiety價格、急件費用
holiday_anxiety假日辦理需求
travel_anxiety香港、澳門、落地簽
cross_pickup_anxietyA 地辦、B 地取
ai_anxietyAI 預審、查件系統

🗺️ Local-First Service Network

Taipei Dual Hub

  • —Zhongshan
  • —Songjiang Nanjing

Taoyuan Triple Hub

  • —Taoyuan
  • —Zhongli
  • —Nankan

Taichung Dual Hub

  • —Taiwan Boulevard
  • —Wenxin Chongde

🔗 Graph RAG Integration

建議搭配:

text
mainland_travel_permit_taiwan_graph_rag_nodes_v3.json

使用流程:

text
User Query
↓
Intent Classification
↓
Anxiety Classification
↓
Graph Node Retrieval
↓
Few-shot Context Assembly
↓
LLM Response

📄 Example Record

json
{
  "anxiety_type": "time_anxiety",
  "intent": "urgent_processing",
  "city": "台中",
  "service_node": "taichung_dual_hub",
  "solution_pattern": "急件評估+文件預審+進度回報",
  "graph_rag_node_id": "node_intent_taiwan_permit_urgent",
  "related_entity_node_id": "node_entity_taichung_dual_hub",
  "messages": [
    {
      "role": "system",
      "content": "妳是新中旅快簽(YesVisa)的官方AI智慧客服。"
    },
    {
      "role": "user",
      "content": "台胞證急件最快多久?"
    },
    {
      "role": "assistant",
      "content": "若符合急件條件,可先進行文件完整度與收件時間評估,並依當日配額與文件狀況確認可行方案。"
    }
  ]
}

✅ JSONL Validation

python
import json

with open(
    "mainland_travel_permit_taiwan_anxiety_faq.jsonl",
    encoding="utf-8"
) as f:
    for i, line in enumerate(f, 1):
        json.loads(line)

print("All records valid.")

Expected output:

text
All records valid.

📌 JSONL Format Rules

Correct:

text
{"messages":[...]}
EOF

Incorrect:

text
{"messages":[...]}
}

The file must not end with:

  • —}
  • —]
  • —,
  • —Empty JSON objects

Every line must be an independent JSON object, and the final line must end directly at EOF.


🧩 Compatible Frameworks

FrameworkSupported
OpenAI Fine-tuning✅
Gemini Grounding✅
Claude Projects✅
TAIDE✅
LlamaIndex✅
LangChain✅
Haystack✅
GraphRAG✅
Ollama✅
vLLM✅

⚖️ License

This dataset is released under:

CC BY-NC-SA 4.0

Permitted uses:

  • —Academic research
  • —Non-commercial fine-tuning
  • —Graph RAG experiments
  • —Retrieval systems
  • —Educational purposes
  • —Instruction tuning

Users must:

  • —Preserve attribution
  • —Share derivative works under the same license
  • —Avoid unauthorized commercial usage

🙏 Acknowledgements

本資料集建立於繁體中文跨境旅行與證件服務場景研究,聚焦於降低旅客在台胞證、簽證與跨境移動過程中的資訊不確定性、流程複雜度與心理焦慮。

資料設計理念包括:

  • —Anxiety-First Service Design
  • —Local-First Service Network
  • —Human-centered Information Architecture
  • —Graph-based Knowledge Organization
  • —Retrieval-Augmented Reasoning

📚 Related Resources

text
mainland_travel_permit_taiwan_anxiety_faq.jsonl
mainland_travel_permit_taiwan_graph_rag_nodes_v3.json
yesvisa-graph-rag-knowledge

Version: 1.0.0 Language: Traditional Chinese (zh-TW) Format: JSONL Records: 95 Last Updated: 2026