Sorry for the delay! I assume you refer to the component variable I select in the beginning. The component variable is a structured field in the original data set that refers to the car part that is the main issue or cause of accident. I select it as a categorical variable so that I can use it in the VTA project for supervised learning purposes in the Category node. I explain the usage of the Category node around 4:15 in the video. In VTA, the software is able to automatically create classification rules for variables I select (in this case I selected the component variable). The classification rules learn keywords from the text corpus that optimize the classification accuracy between different car components. I recently analyzed also a newer set of data from NHTSA (available from www.nhtsa.gov/nhtsa-datasets-and-apis), the the component variable name has changed to compdesc.
Hi, i would like to check what is the component section that you have seleccted as well in the video ?
Sorry for the delay! I assume you refer to the component variable I select in the beginning. The component variable is a structured field in the original data set that refers to the car part that is the main issue or cause of accident. I select it as a categorical variable so that I can use it in the VTA project for supervised learning purposes in the Category node. I explain the usage of the Category node around 4:15 in the video. In VTA, the software is able to automatically create classification rules for variables I select (in this case I selected the component variable). The classification rules learn keywords from the text corpus that optimize the classification accuracy between different car components.
I recently analyzed also a newer set of data from NHTSA (available from www.nhtsa.gov/nhtsa-datasets-and-apis), the the component variable name has changed to compdesc.