Compilation of Malay criminological terms from online news

List of words was gathered through the process of word searching related to each crime category. From the seven categories of crime, a total of 724 crime terminologies were collected. Following the conversion of words into Malay, a total of 521 crime words were left,..the dataset pre-processing was applied to the original dataset. By applying an unsupervised method of filtering using StringToWordVector in WEKA, each word in the text was converted into its attribute. This led to an increase in the total number of attributes. The applied stop words then filtered the attributes by matching the same words to the existing attributes. This pre-processing phase helped obtain attributes in the training and test datasets. A list of 500 most relevant attributes in the content of the output was finalized.

本数据集通过针对各类犯罪类别的词汇检索流程收集词汇列表。从7类犯罪类别中共收集到724条犯罪术语。在将词汇转换为马来语后，最终剩余521条犯罪词汇；随后对原始数据集实施预处理流程。通过使用WEKA中的StringToWordVector工具应用无监督过滤方法，将文本中的每个词汇转换为对应的属性，此举使得属性总数量有所提升。随后通过将词汇与现有属性进行匹配的方式，借助停用词（stop words）对属性完成过滤。该预处理阶段助力获取训练集与测试集所需的属性，最终确定了输出内容中包含的500个最相关属性列表。