SpElementEx:根据ISO-Space规范自动使用空间元素注释文本的工具

  • m0_207279
    了解作者
  • 69.7MB
    文件大小
  • zip
    文件格式
  • 0
    收藏次数
  • VIP专享
    资源类型
  • 0
    下载次数
  • 2022-06-15 03:52
    上传日期
SpElementEx SpElementEx:空间元素提取系统 遵循ISO-Space(Pustejovsky et al。,2013)标注规范的序列标注系统,用于自动标注文本中的空间元素类型和角色信息。 SpElementEx工具已用Java编写,并作为免费软件发布。 先决条件: 输入数据必须为xml格式。 该工具使用的外部库是和 。 在运行该工具之前,必须下载这些库,并且必须使用库jar文件的路径设置Java类路径。 用法: 训练和开发新的空间元素类型提取模型,并使用新开发的模型对测试数据进行空间元素类型注释。 java main.java.spelementex.Main-注释器类型-train -dev -test 训练和开发新的空间元素角色提取模型,并使用新开发的模型对测试数据进行空间元素角色注释。 java main.java.spelementex.Main
SpElementEx-master.zip
内容介绍
# SpElementEx SpElementEx: Spatial Element Extraction System A sequence labeling system for automatically annotating spatial element type and role information in text, following the ISO-Space (Pustejovsky et al., 2013) annotation specifications. The SpElementEx tool has been written in Java and is released as free software. #### Prerequisites: 1) Input data must be in `xml` format. 2) External libraries used by the tool are [Apache Commons IO v2.4](https://commons.apache.org/proper/commons-io/download_io.cgi) and [Stanford CoreNLP](http://nlp.stanford.edu/software/corenlp.shtml#Download). Before running the tool, these libraries must be downloaded and the Java classpath must be set with the paths to the library `jar` files. ### Usage: 1) To train and develop a new spatial element type extraction model, and annotate test data with spatial element types using the newly developed model. java main.java.spelementex.Main -annotators type -train <YOUR TRAIN DIRECTORY> -dev <YOUR DEVELOPMENT DIRECTORY> -test <YOUR TEST DIRECTORY> 2) To train and develop a new spatial element role extraction model, and annotate test data with spatial element roles using the newly developed model. java main.java.spelementex.Main -annotators role -train <YOUR TRAIN DIRECTORY> -dev <YOUR DEVELOPMENT DIRECTORY> -test <YOUR TEST DIRECTORY> `Please note: In this usage, the system requires the input training, development and test data to already have spatial element type annotations. If the input data does not have spatial element type annotations, then follow Usage 3 of the tool below.` 3) To train and develop a new spatial element type and role extraction model, and annotate test data with spatial element types and roles using the newly developed model. java main.java.spelementex.Main -annotators type,role -train <YOUR TRAIN DIRECTORY> -dev <YOUR DEVELOPMENT DIRECTORY> -test <YOUR TEST DIRECTORY> 4) To annotate test data using our pre-trained spatial element type extraction model. java main.java.spelementex.Main -annotators type -test <YOUR TEST DIRECTORY> 5) To annotate test data using our pre-trained spatial element role extraction model. java main.java.spelementex.Main -annotators role -test <YOUR TEST DIRECTORY> `Please note: In this usage, the system requires the input test data to already have spatial element type annotations. If the input test data does not have spatial element type annotations, then follow Usage 6 of the tool below.` 6) To annotate test data using our pre-trained spatial element type and role extraction models. java main.java.spelementex.Main -annotators type,role -test <YOUR TEST DIRECTORY> #### About Sequence Labeling in SpElementEx: SpElementEx employs [CRF++](https://taku910.github.io/crfpp/), a sequence labeling tool based on conditional random fields, for learning and annotating the spatial element information. It is included in SpElementEx's `main\resources\crfpp\` folder. Since [CRF++](https://taku910.github.io/crfpp/) requires feature template files for training a model, files `type-template.txt` and `role-template.txt` in the `main\resources\templates\` folder are created as the features' templates for type and role extraction, respectively.
评论
    相关推荐