| 基于随机森林算法的潜在高价值专利预测方法研究 |
| |
| 引用本文： | 王思培,韩涛.基于随机森林算法的潜在高价值专利预测方法研究[J].情报科学,2020,38(5):120-125. |
| |
| 作者姓名： | 王思培 韩涛 |
| |
| 作者单位： | 中国科学院文献情报中心,北京100190;中国科学院大学经济与管理学院图书情报与档案管理系,北京100190;中国科学院文献情报中心,北京100190;中国科学院大学经济与管理学院图书情报与档案管理系,北京100190 |
| |
| 摘 要： | 【目的/意义】为了支持高价值专利培育工作开展,面向潜在高价值专利预测的需求,提出基于随机森林算法的潜在高价值专利预测方法。【方法/过程】梳理现有研究,选择用于潜在高价值专利预测的指标,构建基于随机森林算法的潜在高价值专利预测模型。使用“语音信号识别”领域的19647条专利进行实证分析,模型预测准确率达96.01%。【结果/结论】目前适于从海量早期申请中发掘潜在高价值专利的方法研究较少,本方法能够在专利申请早期发挥作用,同时具有预测准确率高、处理数据量大、模型可解释性好的优点。
|
| 关 键 词： | 随机森林 高价值专利 专利价值 价值预测 |
Prediction Method of Potential High-value Patents Based on Random ForestAlgorithm |
| |
| WANG Si-pei,HAN Tao.Prediction Method of Potential High-value Patents Based on Random ForestAlgorithm[J].Information Science,2020,38(5):120-125. |
| |
| Authors: | WANG Si-pei HAN Tao |
| |
| Affiliation: | (National Science Library,Chinese Academy of Sciences,Beijing 100190,China;Department of Library,Informaion and Archives Management,School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100190,China) |
| |
| Abstract: | 【Purpose/significance】In order to support the cultivation of high-value patents and meet the demand of the pre diction of high-value patents,a prediction method of high-value patents based on random forest algorithm was proposed.【Method/process】The indexes used for the prediction of potential high-value patents are selected and a prediction method based on Random Forest Algorithm is proposed.Empirical analysis is conducted on 19,647 patents in the field of"voice sig nal recognition",and the prediction accuracy reaches 96.01%.【Result/conclusion】At present,there are few studies on the method suitable for mining potentially high-value patents from massive early applications.This method can play a role in the early application of patents,and it has the advantages of high prediction accuracy,large amount of data processing and good model interpretability. |
| |
| Keywords: | Random Forest Algorithm high value patent patent value value prediction |
| 本文献已被 维普 万方数据 等数据库收录！ |
| 点击此处可从《情报科学》浏览原始摘要信息 |
|
点击此处可从《情报科学》下载全文 |
|
