Author: Alistair Martin; Jama Nateqi; Stefanie Gruarin; Nicolas Munsch; Isselmou Abdarahmane; Bernhard Knapp
Title: An artificial intelligence-based first-line defence against COVID-19: digitally screening citizens for risks via a chatbot Document date: 2020_3_26
ID: 52nw9gxq_33
Snippet: In the most simplistic approach (SF-DIST), we calculated the distance in space between the patient and each of the four diseases, each of which can also be seen as a point in the 10-dimensional symptom space. Normalisation yields the respective probabilities. In the second approach, the same procedure is used, but the distance in each dimension is scaled by the respective standard deviation of each symptom across all diseases (SF-SD). In the thir.....
Document: In the most simplistic approach (SF-DIST), we calculated the distance in space between the patient and each of the four diseases, each of which can also be seen as a point in the 10-dimensional symptom space. Normalisation yields the respective probabilities. In the second approach, the same procedure is used, but the distance in each dimension is scaled by the respective standard deviation of each symptom across all diseases (SF-SD). In the third approach, the distance in each dimension is scaled by the first principal component of a matrix consisting of all symptoms across all diseases (SF-PCA). Lastly, we interpreted the points as vectors and calculated the cosine similarity between the cases and diseases (SF-COS).
Search related documents:
Co phrase search for related documents- cosine similarity and principal component: 1
- cosine similarity and second approach: 1
- cosine similarity and space distance: 1
- disease case and principal component: 1
- matrix principal component and principal component: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
- principal component and space distance: 1, 2
Co phrase search for related documents, hyperlinks ordered by date