基于逻辑回归判断模型的图书馆文献信息自动检索方法
Automatic Retrieval Method of Library Literature Information Based on Logistic Regression Judgment Model
王贺滨1
作者信息
- 1. 郑州工业应用技术学院,河南新郑 451100
- 折叠
摘要
由于传统方法在图书馆文献信息智能检索中应用效果不佳,不仅平均正确率(Mean Average Precision,MAP)较低,而且衡量搜索引擎算法的指标(Discounted Cumulative Gain,DCG)具有较低的归一化止损值,为此提出应用数据挖掘的图书馆文献信息智能检索方法.首先,利用数据挖掘技术估测出嵌入的检索词条与检索文本的相关性;其次,考虑用户反馈构建逻辑回归判断模型;最后,根据逻辑回归判断模型结合最佳权重实现信息检索.实验证明,设计方法MAP值为0.95～0.98,DCG归一化止损值在0.86以上,证明应用设计方法可更好地进行图书馆文献信息检索,应用效果更好.
Abstract
Due to the poor application effect of traditional methods in intelligent retrieval of library literature information,not only is the Mean Average Precision(MAP)relatively low,but also the Discounted Cumulative Gain(DCG),which measures search engine algorithms,has a low normalized stop loss value.Therefore,a library literature information intelligent retrieval method using data mining is proposed.Firstly,using data mining techniques to estimate the correlation between embedded search terms and search text;Secondly,consider user feedback to construct a logistic regression judgment model;Finally,based on logistic regression analysis,the model is combined with the optimal weights to achieve information retrieval.Experimental results have shown that the MAP value of the design method ranges from 0.95 to 0.98,and the normalized stop loss value of DCG is above 0.86,demonstrating that the application of the design method can better perform library literature information retrieval and achieve better application results.
关键词
逻辑回归/图书馆文献信息/智能检索/平均正确率Key words
logistic regression/library literature information/intelligent retrieval/average accuracy引用本文复制引用
出版年
2024
NETL
NSTL
万方数据