Questions tagged [time-series]
Time series are data observed over time (either in continuous time or at discrete time periods).
1,506
questions
0
votes
0answers
4 views
How to predict a certain time span into the future with recurrent neural networks in Keras
I have the following code for time series predictions with RNNs and I would like to know whether for the testing I predict one day in advance:
...
2
votes
1answer
39 views
Google's Bayesian Structural Time-Series
I am attempting to get my head around Google's Causal Impact paper, which isn't completely clear to me.
In the methodology part of the paper, the authors say: "The framework of our model allows ...
1
vote
1answer
24 views
How many features should be there in a dataset to apply any feature selection method?
I am working on a time series, regression problem, where I have 10 features and 180 observations. I would like to understand what the minimum number of features should be in a dataset to use feature ...
1
vote
0answers
16 views
oversampling multivariate time series data
For some classification needs. I have multivariate time series data composed from 4 stelite images in form of (145521 pixels, 4 dates, 2 bands) I made a classification with tempCNN to classify the ...
1
vote
0answers
18 views
Methods to combine datasets from different time periods
Consider a multivariate time series forecasting task where I have two datasets A and B. A goes from 1960 to 2020 and B goes from 2010 to 2020. There is a feature f ...
0
votes
0answers
19 views
Dividing a data set into segments with consistent inner behavior, using segmentation algorithms and metrics for consistency
Context of the problem:
I have signal data which was recorded in a software system and which shows the runtime of multiple processes over time. In total there are more than 900 processes each having ...
0
votes
0answers
11 views
Does this make data leakage in time series? # need help for understanding time series data
Does this make data leakage in time series? I already read this, data leakage when scaling time series
Data leakage is when information from outside the training dataset is used to create the model.
...
0
votes
0answers
8 views
Synchronizing timestamps between multiple sources of time-series data
I have time-series data coming from multiple sensors. The data from each sensor arrives at a "roughly known" interval (i.e. once every 6 hours for example), but the timestamps across the ...
0
votes
1answer
12 views
Serial time conversion in python
I have an array of data "SerTime" which is the sequential time in days, from the start of the series(2016-2018). I am not sure if this is serial time or not. I want to aggregate the data ...
0
votes
0answers
10 views
Why should I use AIC (Akaike information criterion) instead of a metric like RMSE to find the best model?
I have used this AIC metric as a way to find the best SARIMAX model using a grid search to find the values for p,d,q and P,D,Q. I did that because I saw a example of it, but in the end my RMSE result ...
0
votes
0answers
9 views
How can I train a LSTM with different time series of same process?
I have multiple time series dataset of the same process (e.g: sensor collecting humidity in a manufacturing process which last 2 hours) and would like to train a LSTM model to make forecast based on ...
0
votes
0answers
15 views
Removing trend and seasonality does not seem to result in a stationary time series?
I have some sales data, that I want to do time series analysis on. On the plot there are clear trend and seasonality visible.
To test whether a series is stationary I have created a function that ...
0
votes
0answers
6 views
How to achieve windowing and shuffling in tensorflow model of two inputs and two outputs?
I'm analysing daily power consumption of a washing machine. My dataset consists of 4 columns namely:
Unix timestamp
aggregate power consumption by all devices in the house
Power consumption by ...
0
votes
0answers
8 views
How to apply one-to-many LSTM using Keras?
I am finding it difficult to wrap my head around the one-to-many approach using Keras LSTM block.
I have 7 input parameters, using which I need to predict a sequence of length 650.
I referred to LSTM ...
2
votes
0answers
24 views
Models for Long-Term Time-Series Forecasting and Pattern Recognition
I'm trying to find a solution for long-term electricity hourly prices forecasting. Explaining simply, I have some data from 2018 - 2021 containing Demand, Renewable Generation, Hydropower Generation, ...
0
votes
0answers
12 views
What is the best way to create input data samples using in XGBoost for predicting number of next days that customer will come back to store
I'm building the tree-based model like a XGBoost to solve the problem about customer purchase cycle.
And I think, I will build 2 models which one is predicting the customer will come back to store in ...
1
vote
0answers
14 views
Making use of several time series in one LSTM model
I am working on a case where I want to do a multivariate and multi-step time series forecasting.
I have hourly data that measures temperature at approximately 500 different devices. (the devices have ...
1
vote
0answers
13 views
Scipy.fft.fft throws Recursion error "maximum depth exceeded"
N = 4096
# sample spacing
T = 1.0 / 2048.0
yf = fft(sig)
xf = fftfreq(N, T)[:N//2]
This code is from scipy tutorials. I was trying this out on the ...
1
vote
0answers
15 views
estimating frequency in timeseries
I have a statistical question that is difficult to explain, but I try. I appreciate it if you could give me some hints on this issue. My question is: Is there any statistical test or analysis that can ...
0
votes
0answers
17 views
Neural network with variable number of inputs
I have the following problem:
I have a set of time-stamped articles, and labels for particular instants of time. I want to train a neural network such that it can learn which articles to take in as ...
0
votes
0answers
7 views
How to inference of time series with RNN(like LSTM, GRU etc)
Say I am doing a time series prediction which predict some value for next time step with past T inputs from historical inputs. Say I am using a RNN module like LSTM or GRU.
In trainning/validation, I ...
0
votes
0answers
22 views
Improving accuracy of 2D CNN with time series classification
After somewhat extensive optimization of hyperparameters, my test accuracy remains at around 70 %. I have tried techniques to augment time series but they only make things worse.
Unlike image ...
1
vote
0answers
13 views
Time Series Forecasting with LSTMs in keras - convergence problem
I am trying to forecast a time series with multivariate input and multi output (multi step forecast). Since some of my input features are known for future time steps, wheras others are not, naturally ...
1
vote
1answer
20 views
Developing a deep learning hybrid architecture for a particular problem is a highly complicated task [closed]
I am currently conducting research on application of deep learning (sensor signal recognition). I spent about a year and a half sifting through the literature and discovered some research patterns. To ...
1
vote
1answer
51 views
Cross-validation split for modelling data with timeseries behavior
Background: I have a dataset that is generated every month (it is similar with card data that contains card demography and transactions every month and new accounts can be added in the middle of data ...
2
votes
1answer
40 views
How create a representative small subset from a huge dataset, for local development?
I have a time series problem and the dataset I'm using is rather huge. Around 100GB. For local development I'm trying to subset this into a very small batch around 50MB, just to make sure unit tests ...
0
votes
0answers
8 views
FFNN vs. RNN for Regressing Physical Sensor Timeseries Data
I'm trying to build a network to regress data from one sensor to another. The target sensor is a scalar time series and the feature sensor can be either a scalar or vector time series. Both timeseries ...
0
votes
0answers
7 views
Predictions time series
I am very new to time-serieses
I know that we can use LSTM for time-series data
something like this
...
0
votes
0answers
9 views
Calculating all near-tangential lines for inflection points of a time series
I have been thinking about this problem for a while and I'm curious if anyone knows of a good paper on this, has any ideas for algorithms or improvements to the framework.
The task is to store the ...
0
votes
0answers
17 views
Time series Forecasting without consistent timestamps
I am currently working on a time series forecasting model with a dataset that does not have consistent timestamps i.e. one row every 60 seconds. Is it possible to train an accurate model with this ...
1
vote
0answers
7 views
Multiple Time Series Impact
I have a marketing business question where the objective is to learn from my historical data to deduce the best marketing strategy.
Input (Leading Indicator)= For each year I have multiple monthly ...
0
votes
0answers
16 views
Improve accuracy on LSTM - Multiclass Classification problem
Problem Description
I need to build a model which solves the following problem. I have a sequence (let's say size=n) of integers (arrivals) , which looks like this 0,0,1,5,2,...,4,8,6 , and I want to ...
0
votes
0answers
18 views
Recommended ways of splitting train/test in time series in neural networks
For time series, we cannot just split the data and then shuffle it because the training and test sets would have high degree of correlation so we have to split it first and then shuffle each of the ...
0
votes
1answer
20 views
Binary Time Series Forecasting
I am working with daily binary time series forecast as follows:
The target : purchase decision (0: not purchase, 1 purchase
Features: day, weekday, promotion, holiday,....
The objective is trying to ...
1
vote
1answer
27 views
Multivariate Time Series Forecasting using advanced machine learning models
Apologies if this is not the right forum for asking this question. But I have tried other avenues but haven't gotten a satisfactory response. So, finally posting it here.
I've been exploring more ...
0
votes
1answer
52 views
"Up or down but not sideways" bimodal time series prediction - what is the best way to model it?
Say I have a time series (e.g. bitcoin price). I want to predict tomorrow's price, specifically tomorrow's % change in price from today. Let's say this is gaussian distributed, with the mean at 0%.
If ...
0
votes
0answers
22 views
Dense Keras network returns constant output, even for very simple models
I am trying to use keras dense neural networks to forecast some time series.
When fitting my model on complex real datasets, my model converges toward a constant output, i.e. whatever the input, the ...
0
votes
0answers
51 views
Is it always beneficial to use return_sequences=True for time series prediction with RNN?
I roughly understand what return_sequences=True does when being used for time series prediction with RNN (each RNN cell outputs its hidden state). Now my question ...
0
votes
0answers
33 views
Imaging multivariate time series for 2D CNN classification
I have multivariate time series data in the shape of (batches, timesteps, features). So, for 10 samples with 20 timesteps and 4 features, my dataset shape is (10,20,4). I have been using this data for ...
0
votes
1answer
13 views
What are some deep learning models use in timeseries forecasting that include context from covariates?
I was going through the literature for time-series forecasting using DL and all the methods I read about only use the variable of interest at previous timesteps to predict the same variable at time ...
0
votes
0answers
6 views
Is it possible to create a single time series forecasting model to encompass several subseries?
Let's say that I have a univariate time series that measures aggregate sales across all of my company's customers, with daily frequency for a whole year.
Using this time series as my dataset, I ...
2
votes
0answers
23 views
Why the LSTM on Keras does not work correctly when it is necessary to predict several steps forward
I used AirPassenger Dataset. And based on several previous values(for examples 20) I want to predict several(3 or 5) steps in future.
Like
X -> y
[10,20,30,....200]->[210,220,230]
[20,30,40,.......
0
votes
0answers
149 views
Grouped Time Series forecasting with scikit-hts
I am trying to forecast sales for multiple time series I took from kaggle's Store item demand forecasting challenge. It consists of a long format time series for 10 stores and 50 items resulting in ...
1
vote
2answers
66 views
LSTM model, poor performance
I have been working on a project on the demand for a product.
I am using data from 2016 to train the LSTM model.
The architecture is as follows:
...
0
votes
0answers
9 views
Conceptual question - is it correct to use categorical variables such as day, month, year as a fixed sequence input in LSTM?
I am working on a problem where I have to try to predict the dependent variable (continuous) every hour based on hourly temperature (the single continuous variable in predictor space), along with 4 ...
1
vote
1answer
36 views
When to tune hyperparameters in deep learning
I am currently playing around with different CNN and LSTM model architectures for my multivariate time series classification problem.
I can achieve validation accuracy of better than 50 %. I would ...
0
votes
0answers
11 views
How to find the lag that shows highest correlation between two time series variables?
I have two variables in my time series data (X1,X2). I need to find the correlation between these two variables at different lags and identify the lag that shows the highest correlation. For e.g. I ...
0
votes
1answer
11 views
Regression prediction for HVAC unit Best way to utilize available data?
I am starting to investigate machine learning applications for HVAC at the commercial level. I am an HVAC controls person by trade that has recently taken some basic courses on Machine learning and ...
0
votes
1answer
34 views
which statistical parameters are more useful to detect anomalies and outlier? mean max min var?
This time series contains some time frame which each of them are 8K (frequencies)*151 (time samples) in 0.5 sec [overall 1.2288 millions samples per half a second)
I need to find anomalous based on ...
0
votes
0answers
13 views
Find part of the signal that causes most of the noise
We are given around 20 time series. Each one of them is how many products were sold every day (and they have different volume, different mean and std). The main goal is to predict the total sum of ...
