Title: RESEARCH COMMUNICATIONS OF THE 28th ECVIM-CA CONGRESS Document date: 2018_12_19
ID: r79h9yzz_914
Snippet: Multiple tests were conducted, the training data varying between 8 and 15 patients. The experimental results revealed a confidence over 95% for both types of ANN's, regardless of the type of insulin. Accuracy performance obtained on training and validation datasets proved that the use of the adopted neural network vector is relevant to the dynamics of the glycemic evolution in feline diabetes mellitus. Results suggest predictive modeling is feasi.....
Document: Multiple tests were conducted, the training data varying between 8 and 15 patients. The experimental results revealed a confidence over 95% for both types of ANN's, regardless of the type of insulin. Accuracy performance obtained on training and validation datasets proved that the use of the adopted neural network vector is relevant to the dynamics of the glycemic evolution in feline diabetes mellitus. Results suggest predictive modeling is feasible for establishing and adjusting insulin dose, improve treatment effectiveness, increase diabetes remission rate and consequently obtain a better quality of life of diabetic cats.
Search related documents:
Co phrase search for related documents- diabetic cat and feline diabetes: 1
- diabetic cat and feline diabetes mellitus: 1
Co phrase search for related documents, hyperlinks ordered by date