SOMson -- Sonification of Multidimensional Data in Kohonen Maps

Kohonen地图，也称为自组织映射（SOMs），是一种神经网络，可以将高维特征空间可视化在低维地图上。虽然SOM是数据检查和探索的绝佳工具，但它们本质上会导致信息的丢失。底层数据的可视化不够完整，因此无法提供整体图片。因此，我们建议使用SOMson作为数据增强技术，以交互式的方式将底层数据声音化。声音化可以增加SOM同时提供的信息量。我们没有进行用户研究，而是提供了一个交互式的在线示例，让读者自己探索SOMson。文章讨论了SOMson的优点、缺点和前景。

SOMson: An Interactive Sonification of Self-Organizing Maps for Data Augmentation

The paper proposes SOMson, an interactive sonification technique to augment the information provided by Self-Organizing Maps (SOMs) for data exploration. SOMson converts the data points in the SOM into sound, allowing users to simultaneously explore the visual and auditory representations of the data.

The paper presents an interactive online example of SOMson for readers to explore. The authors discuss the strengths, weaknesses, and prospects of SOMson for data exploration. The paper also highlights the potential of SOMson for assisting individuals with visual impairments. However, no user study or empirical evaluation of SOMson is presented in this paper.

Related studies in this field include: 'Sonification of Self-Organizing Maps for Temporal Data Exploration' by P. Knees et al., and 'SonicMaps: Exploring Self-Organizing Maps with Sound' by M. M. Wanderley et al.