predict probabilities using LSTM

6 次查看（过去 30 天）
Gadelhag M Omar Mohmed
回答： Ankit Kumar 2019-10-21
Hi all
I am trying to create a predict probabilities model using LSTM for predicting 5 different classes. I am just wondering is it possible to have the output of the prediction model using LSTM as a vector of belonging values to each class, not as a crisp value? In other words, if I have to predict 5 different classes, when I use the prediction process, I want to have the results as a vector of belonging degrees to each output class for each instance such as:
pred_out= [0.025, 0.8, 0.00058, 0.225. 0.07]. So, the highest value for class number 2. Any ideas, please.
Thanks in advance
Gad

请先登录，再回答此问题。

采纳的回答

Ankit Kumar
Ankit Kumar 2019-10-21
Hi Gadelhag,
You can use
[YPred, scores] = classify(net,sequences)
Here scores give you the probabilities of different classes that you want.
For more about classify function refer,
https://in.mathworks.com/help/deeplearning/ref/classify.html#d117e99562

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Image Data Workflows 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by