公共财政绩效评价指标 信息数据
收藏资源简介：
本数据集的核心价值在于将公共财政绩效评价项目中的评价指标文本信息进行了提取、特征化、识别、标准化、格式化等处理，从而提高了数据的可用性和分析效率。具体应用场景包括但不限于：（1）AI模型的开发与训练：本数据集可以作为指标体系生成、绩效趋势预测等AI模型开发和训练的高质量的基础数据，从而提高AI模型的训练效率和准确性。（2）公共财政绩效指标体系构建：由于本数据集涵盖的评价指标信息的全面性，可为政府部门构建和完善预算绩效指标体系提供有效的借鉴和参考。1.数据采集：从公司的公共财政绩效评价历史项目库的报告中，采集与项目基本信息和评价指标相关的内容，格式包括doc、docx、pdf和图像。记录被采集报告的项目名称和采集的时间。 2.文本提取：使用Apache POI库对doc、docx格式进行解析，使用PDFMiner工具对pdf格式进行解析，提取其中的文本内容。使用tesseract OCR光学字符识别工具将图像中的文本转换为机器可读的文本格式。 3.数据预处理：1）去除提取出的文本中与项目基本信息和评价指标无关的字符。2）对文本内容进行分词处理，以便于提取关键词和短语。3）删除常见的、意义不大的词汇。 4.特征提取：使用TF-IDF（词频-逆文档频率）算法提取文本中的关键词。 5.指标识别：使用自然语言处理（NLP）对评价指标进行分类识别。 6.人工检查：通过公司专业人员对NLP识别的评价指标进行审核，确保识别结果的准确性和可靠性。 7.数据标准化：将识别出的评价指标按照统一的格式进行归一化处理，如对评价指标的单位进行统一等。将提取的评价指标信息格式化为统一的csv数据结构。记录数据完成处理的时间，并赋予唯一编号。
The core value of this dataset lies in processing the text information of evaluation indicators in public financial performance evaluation projects via extraction, feature engineering, recognition, standardization and formatting, thereby enhancing data usability and analytical efficiency. Specific application scenarios include but are not limited to: (1) Development and training of AI models: This dataset can serve as high-quality basic data for developing and training AI models such as indicator system generation and performance trend prediction, thereby improving the training efficiency and accuracy of AI models. (2) Construction of public financial performance indicator systems: Given the comprehensiveness of the evaluation indicator information covered by this dataset, it can provide effective references for government departments to construct and improve budget performance indicator systems. 1. Data Collection: Collect content related to basic project information and evaluation indicators from the reports stored in the company's historical project database of public financial performance evaluations. The supported collection formats include doc, docx, pdf and images. Record the project name of the collected report and the collection time. 2. Text Extraction: Use the Apache POI library to parse doc and docx files, employ the PDFMiner tool to parse pdf files, and extract the embedded text content. Use the Tesseract OCR (Optical Character Recognition) tool to convert text within images into machine-readable text format. 3. Data Preprocessing: 1) Remove characters irrelevant to basic project information and evaluation indicators from the extracted text; 2) Conduct word segmentation on the text content to facilitate the extraction of keywords and phrases; 3) Remove common stop words that carry minimal semantic meaning. 4. Feature Extraction: Use the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm to extract keywords from the processed text. 5. Indicator Recognition: Utilize natural language processing (NLP) technologies to classify and recognize the evaluation indicators. 6. Manual Inspection: Have professional personnel from the company review the evaluation indicators identified by the NLP model to ensure the accuracy and reliability of the recognition results. 7. Data Standardization: Normalize the identified evaluation indicators into a unified format, such as unifying the units of the evaluation indicators. Format the extracted evaluation indicator information into a unified CSV (Comma-Separated Values) data structure. Record the data processing completion time and assign a unique serial number to each dataset entry.
