V
V
Vladislav Satarov2016-03-01 17:44:14
Neural networks
Vladislav Satarov, 2016-03-01 17:44:14

What type of neural networks to use to predict data?

There is depersonalized data for several years about the object.
The data is tied to time. (dd/mm/yy number)
The format of this file is xml or xls .
It is necessary to predict the result for a future date
. Which neural network to choose and what method should be used to train it?

Answer the question

In order to leave comments, you need to log in

3 answer(s)
Y
Yane, 2016-03-10
@satar_V_A

Recurrent networks have short-term memory (i.e., they can infer from several previous input values, not just one previous value). The most famous types of them now are LSTM and GRU. Used, for example, to predict the next word in a sentence based on the previous ones. You can try them in your case, I guess.

N
nirvimel, 2016-03-01
@nirvimel

If the data are quantitative (not categorical) in nature and are presented as numerical values, then predicting future results from the past is a classic regression task and neural networks are not related to this.

L
lomikman, 2016-03-01
@lomikman

you can ask your question here https://www.kaggle.com/
sometimes they are responsible for money

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question