中成药产业链结构文本训练数据
收藏资源简介：
本数据集服务于中成药产业链智能分类与产业图谱构建模型的训练与开发，通过关联企业文本与产业链环节标签，为中医药产业管理与服务提供数据工具。其主要应用于： 供应链管理与采购寻源：赋能医院、连锁药店及医药流通企业，精准识别上游药品生产企业（如特定剂型、临床用途）及下游合规零售终端，优化采购决策与供应链布局。区域产业规划与监管：辅助地方政府与药监部门，分析区域内中成药研发生产企业的技术分布（如内科用药、冲剂生产）以及零售终端的服务能力，为产业规划与市场监管提供数据支持。市场研究与投资分析：支持行业研究机构与投资方，洞察中成药不同细分品类（如肿瘤用药、妇产科用药、特色中药）的市场竞争格局、企业研发动态与零售渠道分布。一、加工前数据说明 本数据集旨在构建用于中成药产业链智能分析的人工智能模型训练语料。在加工前，数据已进行严格的匿名化与去标识化处理。原始企业名称被统一替换为不可逆的规范标识符，并彻底移除所有的个人及商业敏感信息，确保数据完全符合隐私保护与安全合规要求，为模型训练提供了洁净、可靠的输入基础。 二、数据处理规则 数据处理严格遵循 “体系先行、业务匹配、特征抽取” 的核心规则，形成了一套从分类框架构建到最终标签生成的完整流程：1.首先，依据国家药品分类标准及中医药产业链分工，预先定义了以“中成药”为一级节点，按产业链位置划分为“中游：中成药研发生产”和“下游：中成药服务及应用”两个二级节点。并进一步细分为“临床用途”、“剂型”、“医药流通”等三级节点，以及“内科用药”、“冲剂”、“医药零售”、“药店”等四级节点的树状分类体系，为数据加工提供了清晰的产业与技术逻辑框架。2.业务匹配：采用“自动化规则匹配与人工校验相结合”的策略。首先，依托Spark大数据处理框架，对海量企业简介文本进行分布式清洗、分词与关键词匹配，通过预构建的中成药产业语义规则库自动计算并推荐初步分类节点。随后，由具备医药行业知识的标注专家进行审核与最终判定，确保企业被精准归入对应的产业链环节与业务类型。3.特征抽取：在完成业务匹配的同时，从同一段企业简介文本中，系统性地抽取代表其核心产品与技术的关键术语与名词性短语，经过去重与标准化格式化，组合成“正向词”特征串，作为对分类标签的语义补充。 三、加工后数据内容 加工后的数据集为一条条结构化的“文本-标签”数据。每条数据均包含经过脱敏处理的原始企业描述文本，以及与之对应、经人工校验的完整分类标签（一至四级节点）、高度细化的业务特征词（正向词）与产业标签。数据内容全面覆盖了中成药产业链上游的研发生产（涵盖不同剂型与临床用途）和下游的流通服务（涵盖批发、零售、药店等终端）等核心环节，形成了一个分类体系专业、业务特征鲜明、可直接用于中成药产业链分析、生产企业分类、零售终端识别与市场格局研究等模型训练与评估的高质量专用数据集。
This dataset is developed for the training and development of intelligent classification and industrial knowledge graph construction models for the Chinese patent medicine industrial chain. By associating enterprise texts with industrial chain segment labels, it provides data tools for the management and service of the traditional Chinese medicine (TCM) industry. Its main applications are as follows: 1. Supply Chain Management and Sourcing: Empowering hospitals, chain pharmacies and pharmaceutical distribution enterprises to accurately identify upstream pharmaceutical manufacturers (e.g., specific dosage forms, clinical indications) and compliant downstream retail terminals, so as to optimize procurement decisions and supply chain layout. 2. Regional Industrial Planning and Regulation: Assisting local governments and drug regulatory authorities to analyze the technical distribution of Chinese patent medicine R&D and manufacturing enterprises in the region (e.g., internal medicine medications, granule production) and the service capabilities of retail terminals, providing data support for industrial planning and market supervision. 3. Market Research and Investment Analysis: Supporting industry research institutions and investors to gain insights into the market competition landscape, enterprise R&D trends and retail channel distribution of different segmented categories of Chinese patent medicines (e.g., oncology medications, gynecology and obstetrics medications, characteristic traditional Chinese medicines). I. Pre-Processing Data Description This dataset is intended to serve as training corpus for AI models used in intelligent analysis of the Chinese patent medicine industrial chain. Prior to processing, the data has undergone strict anonymization and de-identification processing. Original enterprise names have been uniformly replaced with irreversible standardized identifiers, and all personal and commercial sensitive information has been completely removed, ensuring that the data fully complies with privacy protection and security compliance requirements, providing a clean and reliable input foundation for model training. II. Data Processing Rules The data processing strictly follows the core principles of "system first, business alignment, feature extraction", forming a complete workflow from classification framework construction to final label generation: 1. First, based on national pharmaceutical classification standards and the division of labor in the TCM industrial chain, a tree-like classification system is pre-defined, with "Chinese patent medicine" as the first-level node, divided into two second-level nodes according to industrial chain positions: "Midstream: Chinese Patent Medicine R&D and Manufacturing" and "Downstream: Chinese Patent Medicine Services and Applications". It is further subdivided into third-level nodes such as "clinical indications", "dosage forms", "pharmaceutical distribution", and fourth-level nodes such as "internal medicine medications", "granules", "pharmaceutical retail", "pharmacies", providing a clear industrial and technical logical framework for data processing. 2. Business Alignment: Adopting a strategy of combining automated rule matching and manual verification. First, relying on the Spark big data processing framework, distributed cleaning, word segmentation and keyword matching are performed on massive enterprise profile texts, and preliminary classification nodes are automatically calculated and recommended through a pre-built semantic rule base for the Chinese patent medicine industry. Subsequently, annotation experts with pharmaceutical industry knowledge conduct reviews and final judgments to ensure that enterprises are accurately classified into corresponding industrial chain segments and business types. 3. Feature Extraction: Upon completing business alignment, key terms and noun phrases representing their core products and technologies are systematically extracted from the same enterprise profile text. After deduplication and standardization formatting, they are combined into a "positive terms" feature string, which serves as a semantic supplement to the classification labels. III. Post-Processing Data Content The post-processed dataset consists of structured "text-label" entries. Each entry contains desensitized original enterprise description text, as well as corresponding manually verified complete classification labels (first to fourth-level nodes), highly refined business feature words (positive terms) and industrial labels. The data content comprehensively covers core links of the Chinese patent medicine industrial chain, including upstream R&D and manufacturing (covering different dosage forms and clinical indications) and downstream circulation services (covering terminals such as wholesale, retail, pharmacies, etc.). It forms a high-quality dedicated dataset with professional classification system, distinct business characteristics, which can be directly used for model training and evaluation in Chinese patent medicine industrial chain analysis, manufacturing enterprise classification, retail terminal identification and market landscape research.
