摘要
自然语言具有模糊性和歧义性特点,加大了特征提取难度,为了能够精准提取自然语言特征,提出一种基于模糊关联优化的自然语言特征提取方法。将不确定性自然语言信息利用三元语言表示模型描述,给出一个初始隶属度函数(MF),设定最大化模糊项集支持度和语义可解释性为适应度函数,利用群搜索优化(GSO)算法获取最佳MF,通过优化后的模糊关联规则算法挖掘自然语言信息。在注意力机制中加入生成函数和限制函数,改进传统注意力机制,基于改进后的注意力机制完成自然语言特征提取。仿真结果表明,所提方法可以获取高精度与高覆盖率的自然语言特征提取结果。
The fuzziness and ambiguity of natural language increase the difficulty of feature extraction.To accurately extract natural language features,a method of extracting natural language features based on fuzzy association optimization was proposed.Initially,uncertain natural language information was described by using a ternary language representation model.Then,an initial membership function(MF)was provided.Next,maximizing the support of fuzzy item sets and semantic interpretability were set as the fitness function.Moreover,the Group Search Optimization(GSO)algorithm was adopted to obtain the optimal MF.Furthermore,natural language information was mined by the optimized fuzzy association rule algorithm.After that,the generation function and restriction function were added to the attention mechanism.Meanwhile,the traditional attention mechanism was improved.On this basis,the extraction of natural language features was completed.Simulation results show that the proposed method can achieve high-precision and high-coverage natural language feature extraction.
作者
蓝桂军
李民
LAN Gui-jun;LI Min(College of Mechanical and Electrical Engineering,Kunming University of Science and Technology,Kunming Yunnan 650000,China)
出处
《计算机仿真》
2024年第8期234-237,302,共5页
Computer Simulation
基金
基于模糊关联优化的自然语言特征提取仿真(XYYJ2022C01)。
关键词
模糊关联优化
自然语言
特征提取
限制函数
注意力机制
Fuzzy correlation optimization
Natural language
Feature extraction
Restriction function
Attention mechanism
