Global suicide mortality rates (2000-2019) and bibliographic data

The dataset contains World Bank Suicide mortality rate WDI (world development indicator) (2000-2019) world-wide data in original and processed form. In addition to the statistical data this dataset also contains bibliographic records of articles published on the topic of suicide in relation to individual countries during (2000-2019) in original and processed form. 

The data consists of six archives:

World development indicator suicide mortality rate SH.STA.SUIC.P5. This archive contains suicide mortality rate of 159 countries during the period of 2000-2019 per 100,000 population including males and females as of November, 2023.

Web of science records country and suicide. This archive contains bibliographic records organized by country on the topic of suicide related to that country published during 2000-2019 as of November, 2023.

Suicide mortality rate statistics and keywords. This archive contains processed data of 1 and 2 archives in three files. The 'Countries suicide rates and WOS records' contains organized temporal suicide mortality rate data for each country and each year for males and females including counts of articles on suicide related in that country. The 'words and countries matrix' file contains information about how many times author and paper keywords from suicide related publications were seen in articles associated with each country. This data is organized as matrix in which rows are keywords, columns are countries and cells are counts of the keyword. The 'words and countries pairs' file contains same information only organized as keyword country pairs.

Suicide mortality rate clusters countries keywords titles. This archive contains bibliographic data organized by country clusters. These clusters group countries with similar suicide mortality rate dynamics in males and females shown in two included figures. Each folder of the cluster contains a section with bibliographic records; a section with keywords associated with each country; and a section in which each publication associated with the country has a separate filecontaining its title and keywords.

Suicide keywords embedding data. This archive contains word embedding vectors and metadata learned by recurrent neural network trained to classify countries from suicide related keywords of articles associated with those countries. Folder 'trained with keywords' contains embeddings learned in classifying countries in which training samples are keyword strings of publications. Folder 'trained with titles' contains embeddings learned in classifying countries in which training samples are strings containing titles of publication plus keywords.

Suicide keywords association rule mining. This archive contains files of subsets of keywords frequently mentioned together in suicide related publications. Folder 'Mining in clusters' has frequent keyword itemsets in country clusters. Folder 'Mining in individual countries' has frequent keyword itemsets in countries. Examples of keyword networks connecting clusters and networks connecting countries in individual clusters are included which helps to identify specific and shared keywords by country clusters and by countries in the individual clusters. 

These datasets support a data availability statements for upcoming articles.

本数据集收录世界银行自杀死亡率世界发展指标（World Development Indicator, WDI）2000-2019年的全球原始与处理后数据。除统计数据外，本数据集还包含2000-2019年间发表的、与各国自杀主题相关的文献著录条目，同样提供原始与处理后两种格式。

本数据集共包含六个档案包：
1. 世界发展指标自杀死亡率（SH.STA.SUIC.P5）：该档案包收录截至2023年11月的159个国家2000-2019年每10万人口自杀死亡率数据，涵盖男性与女性群体。
2. 科学网（Web of Science）国家与自杀主题文献：该档案包包含按国家分类的2000-2019年间发表的、与该国自杀相关主题的文献著录条目，数据更新至2023年11月。
3. 自杀死亡率统计与关键词：该档案包包含前两个档案包的处理后数据，分为三个文件：
- 《各国自杀率与科学网文献》（Countries suicide rates and WOS records）：收录按年份整理的各国分性别自杀死亡率时序数据，以及对应国家自杀相关文献的发表总量；
- 《关键词与国家矩阵》（words and countries matrix）：收录各国自杀相关出版物中作者关键词与论文关键词的出现频次信息，数据以矩阵形式组织，行代表关键词、列代表国家，单元格数值为对应关键词的出现次数；
- 《关键词与国家配对表》（words and countries pairs）：包含与上述文件完全一致的信息，仅以关键词-国家配对的形式进行组织。
4. 自杀死亡率聚类：国家、关键词与文献标题：该档案包包含按国家聚类整理的文献数据。聚类以各国分性别自杀死亡率的变化趋势为依据进行分组，配套包含两张可视化图表。每个聚类文件夹包含三个模块：文献著录条目模块、与各国关联的关键词模块，以及每个与该国相关的出版物的独立文件（包含其标题与关键词）。
5. 自杀关键词词嵌入数据：该档案包包含由循环神经网络（recurrent neural network）训练得到的词嵌入（word embedding）向量与元数据，该模型用于从自杀相关出版物的关键词中完成国家分类任务。其中「基于关键词训练」（trained with keywords）文件夹收录以出版物关键词字符串作为训练样本的国家分类任务所学习得到的词嵌入；「基于标题与关键词训练」（trained with titles）文件夹收录以出版物标题与关键词拼接字符串作为训练样本的国家分类任务所学习得到的词嵌入。
6. 自杀关键词关联规则挖掘：该档案包收录自杀相关出版物中高频共现关键词子集的相关文件。其中「聚类内挖掘」（Mining in clusters）文件夹包含各国聚类内的高频关键词项集；「单国内挖掘」（Mining in individual countries）文件夹包含单个国家内的高频关键词项集。文件中还附带连接聚类的关键词网络示例，以及单个聚类内连接各国的关键词网络示例，可用于识别各国聚类及单个聚类内各国特有的与共有的关键词。

本数据集可为后续学术论文的数据可用性声明提供支撑。