Developing a Method for Identifying Instances of Group Generative Interactions in Enterprise Social Media
Developing a Method for Identifying Instances of Group Generative Interactions in Enterprise Social Media
复制标题
DOI：
--
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Elisavet Averkiadi;W. V. Osch;Yuyang Liang
中科院分区：
文献类型：
--
作者：
Elisavet Averkiadi;W. V. Osch;Yuyang Liang
公司对群体生成互动特别感兴趣-通过群体交流产生新想法和解决方案的概念。它们是创新的根源，因此对企业的生存至关重要。企业社交媒体（ESM）提供了一个独特的机会来研究生成群体互动，由于这些平台上的活动是透明的。在这篇正在进行中的研究论文中，我们进行了初步分析，以开发一种可以识别基于esm的生成群体交互实例的方法，其中我们重点区分生成与非生成群体交互。为了做到这一点，我们使用了来自跨国组织的ESM平台的所有小组互动的文本。我们实现了机器学习模型来学习和分类文本为生成或非生成。结果，我们从表现最好的模型中产生了最重要的术语特征。这些特征将有助于我们在未来的研究中理解这些相互作用中发生的讨论的性质。
Companies hold particular interest in group generative interactions - the conception of novel ideas and solutions through group exchanges. They are a root-cause of innovation and thus are important to companies’ survival. Enterprise Social Media (ESM) offer a unique opportunity to study generative group interactions, due to the transparent nature of activities on these platforms. In this research-in-progress paper, we conduct a preliminary analysis to develop a method that could identify the instances of ESM-based generative group interactions, where we focus on distinguishing generative versus non-generative group interactions. To do this, we used the text from all group interactions from an ESM platform of a multinational organization. We implemented machine learning models to learn and classify the text as generative or non-generative. As a result, we produced the top important term features from the best performing model. These features will help us understand the nature of discussions that occur in these interactions in future studies.