Language intensity classification and Neuronal Networks

Kohonen Self-Organizing Maps (SOM) are a particular type of artificial neural network created by Teuvo Kohonen. Their unsupervised learning makes them suitable for application, among other things, to grouping tasks. Occupations have been grouped into this analysis considering the similarity in value of the variables that define this occupation in terms of language proficiency requirements. The 8 variables used were: v1-speaking skills, v2-writing skills, v3-speech clarity, v4-speech recognition, v5-English knowledge, v6-speaking ability, v7-communication with people outsiders, v8-communication with superiors, equals or subordinates. After applying this authomatic classification technique, the SOC occupations are classified according to five distinc groups with decreasing linguistic intensity: Class 1: High linguistic intenisty requirements<br> Class 2: Medium-high linguistic intenisty requirements<br> Class 3: Medium linguistic intenisty requirements<br> Class 4: Medium-low linguistic intenisty requirements<br> Class 5: Low linguistic intenisty requirements科赫农自组织映射（Kohonen Self-Organizing Maps，SOM）是由特沃·科赫农（Teuvo Kohonen）提出的一类特殊人工神经网络。该模型采用无监督学习方式，可适用于聚类任务等多种应用场景。本分析基于职业在语言能力要求维度上的定义变量取值相似度，对职业开展聚类分组。本次分析共使用8项变量，分别为：v1 口语表达能力、v2 书面写作能力、v3 语音清晰度、v4 语音识别能力、v5 英语知识水平、v6 口语运用能力、v7 与外部人员沟通能力、v8 与上级、平级或下级的沟通能力。在应用该自动分类技术后，SOC职业被划分为5个语言强度依次递减的组别，具体如下： 类别1：高语言强度要求 类别2：中高语言强度要求 类别3：中等语言强度要求 类别4：中低语言强度要求 类别5：低语言强度要求