0
$\begingroup$

I am working on LSTM and CNN to solve the time series prediction problem.

I have seen some tutorial examples of time series prediction using CNN-LSTM. But I don't know if it is better than what I predicted using LSTM.

Could using LSTM and CNN together be better than predicting using LSTM alone?

Improve this question
$\endgroup$
5
  • 2
    $\begingroup$ You wrote "I have seen some tutorial examples of time series prediction using CNN-LSTM". Can you please link us to those tutorials? Moreover, you should probably describe your dataset (i.e. is it a sequence of images? or how many data points do you have?). $\endgroup$
    – nbro
    Commented Dec 8, 2020 at 11:24
  • $\begingroup$ Have a look at stats.stackexchange.com/questions/10162/… $\endgroup$
    – kjetil b halvorsen
    Commented Feb 3, 2021 at 1:08
  • $\begingroup$ Similar to stats.stackexchange.com/questions/578508/… although no answers posted. $\endgroup$
    – Brian O'Donnell
    Commented Jul 24, 2022 at 17:02
  • $\begingroup$ Please provide details on the time series. For example, is it rainfall in the Amazon, prices of airfares, or is it something chaotic like stock prices? In addition, is your data univariate or multivariate? $\endgroup$
    – Brian O'Donnell
    Commented Jul 24, 2022 at 17:24
  • $\begingroup$ It could, but it doesn't mean it always will. I guess that's why it can be called "Data Science" problem, because every dataset is different and needs its own little "research". So what I mean is that you have to implement CNN-LSTM model and see for yourself. $\endgroup$
    – GKozinski
    Commented Aug 24, 2022 at 10:26

3 Answers 3

Reset to default
1
$\begingroup$

Many papers have been published on CNN, LSTM, and CNN-LSTM for time series. From the literature and my experience, I conclude that CNN-LSTM outperforms CNN and LSTM models. Here are two relevant papers on stock price time series forecasting:

Wenjie Lu, Jiazheng Li, Yifan Li, Aijun Sun, & Jingyang Wang. (2020). A CNN-LSTM-Based Model to Forecast Stock Prices. Complexity, 2020. https://doi.org/10.1155/2020/6622927

The authors' conclusions:

This paper takes the relevant data of the Shanghai Composite Index as an example to verify the experimental results. The experimental results show that the CNN-LSTM has the highest forecasting accuracy and the best performance compared with the MLP, CNN, RNN, LSTM, and CNN-RNN.

Chen, Y., Fang, R., Liang, T., Sha, Z., Li, S., Yi, Y., Zhou, W., & Song, H. (2021). Stock Price Forecast Based on CNN-BiLSTM-ECA Model. Scientific Programming, 1–20. https://doi.org/10.1155/2021/2446543

The authors' conclusions:

The proposed model is compared with CNN, LSTM, BiLSTM, CNNLSTM, CNN-BiLSTM, BiLSTM-ECA, and CNN-LSTMECA network models on three datasets. The experimental results show that the proposed model has the highest prediction accuracy and the best performance.

Improve this answer
$\endgroup$
0
$\begingroup$

In my opinion, if your dataset has correlation with its neighbors and there is some kind of sequence involved in it, then CNN-LSTM would provide you with much better results than only CNN or LSTM.

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Your answer could be improved with additional supporting information. Please edit to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers in the help center. $\endgroup$
    – Community Bot
    Commented Jul 24, 2022 at 9:44
0
$\begingroup$

CNN can be used as good feature extractor to help LSTM layers train efficiently , normally when your input contains huge number of features (images , audio) and you need to project these features into learnable lower dimension space use CNN before LSTM

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .