The LMF ISO standard provides a large cover of lexical knowledge using a fine structure. However, like most of the electronic dictionaries, the available normalized LMF dictionaries comprise only basic morpho-syntactic and semantic knowledge, such as the meanings of lexical entries through the definitions and the associated examples, and sometimes the indication of the synonyms and antonyms. Other sophisticated knowledge, such as the syntactic behaviors, semantic classes and syntactico-semantic links, which are scarce, requires a high expertise and its adding to dictionaries is expensive. In fact in this paper, we propose an approach of lexical data mining of the widely available textual content associated with the meanings, notably in the normalized LMF dictionaries, in order to perform the self-enrichment of these dictionaries. First, we contribute to the enrichment of the syntactic behaviors by linking them to the suitable meanings. Second, we focus on the enrichment of the meanings of LMF lexical entries with semantic classes based on the Gaston Gross semantic classification. Finally, we establish the syntactico-semantic links based on the results of the syntactic and semantic enrichment processes. The proposed approach has been consolidated by an experimentation carried out on an available normalized LMF dictionary for Arabic language.

中文翻译：

---

基于词法数据挖掘的 LMF 标准化词典自我丰富方法：句法语义知识案例

LMF ISO 标准使用精细结构提供了大量词汇知识。但是，与大多数电子词典一样，可用的规范化 LMF 词典仅包含基本的形态句法和语义知识，例如通过定义和相关示例了解词条的含义，有时还包括同义词和反义词的指示。其他复杂的知识，例如句法行为、语义类和句法-语义链接，这些都是稀缺的，需要很高的专业知识，并且将其添加到词典中的成本很高。事实上，在本文中，我们提出了一种对广泛可用的与含义相关的文本内容进行词法数据挖掘的方法，特别是在规范化的 LMF 词典中，以便对这些词典进行自我丰富。第一的，我们通过将句法行为与合适的含义联系起来，为丰富句法行为做出贡献。其次，我们专注于基于Gaston Gross语义分类的语义类丰富LMF词条的含义。最后，我们根据句法和语义丰富过程的结果建立句法-语义链接。提议的方法已通过对可用的阿拉伯语规范化 LMF 词典进行的实验得到巩固。我们根据句法和语义丰富过程的结果建立句法-语义链接。提议的方法已通过对可用的阿拉伯语规范化 LMF 词典进行的实验得到巩固。我们根据句法和语义丰富过程的结果建立句法-语义链接。提议的方法已通过对可用的阿拉伯语规范化 LMF 词典进行的实验得到巩固。