Train and Deploy Time series forecasting models on Tiyaro using Tiyaro EasyTrain

Train and Deploy Time series forecasting models on Tiyaro using Tiyaro EasyTrain

A quicker and easier way to train and deploy time-series forecasting models in 5 simple steps

Sharvil Mainkar's photo
Sharvil Mainkar
·Aug 5, 2022·

2 min read

Play this article


Time series forecasting is the task of creating a model to predict future values based on previously observed values. Time series forecasting is used in various different industries from retail for forecasting the daily sales volume to agriculture for forecasting weather to guide planning decisions around planting and harvesting. Tiyaro allows an easy way to train and quickly predict your time series using classical and newer machine learning methods.

These are the 5 simple steps to create a time series forecasting API:

Step 1: Gather your time series dataset:

Before we dive into the part of training the model, it is important to have a clean dataset in the correct format for our training time series forecasting models using Tiyaro's EasyTrain. The format required for the dataset can be found over in our Tiyaro Docs.

Step 2: Create an EasyTrain Job:

The quickest way to train a time series forecasting model is by going over the side tab and clicking on Retrain

Screen Shot 2022-08-04 at 11.20.45 PM.png

Screen Shot 2022-08-04 at 11.22.00 PM.png

Screen Shot 2022-08-04 at 11.23.51 PM.png

Select a model class for our case Time Series forecasting.

Screen Shot 2022-08-04 at 11.24.19 PM.png

Step 3: Upload your dataset and fill out training job information:

Screen Shot 2022-08-04 at 11.35.45 PM.png

Select all the models you want to train. EasyTrain takes care of hyperparameter optimization.

Screen Shot 2022-08-04 at 11.36.21 PM.png

Step 4: Start Training your EasyTrain Job:

Screen Shot 2022-08-04 at 11.36.43 PM.png

Step 5: Enjoy your results.

As you can see, we have trained 7-time series forecasting models on our dataset.

Screen Shot 2022-08-04 at 11.37.11 PM.png

Here, we show the time series forecasting Prophet model, we can use the demo tab to see the forecasted values in the future.

Screen Shot 2022-08-04 at 11.37.53 PM.png

The trained model page showcases information about the API as well as the speed and cost of the model.

Screen Shot 2022-08-04 at 11.38.07 PM.png


As we saw it is easy for a user to use Tiyaro EasyTrain to create time series forecasting models for their particular use case. Tiyaro makes it easier for the user to quickly train a wide variety of Time Series Forecasting models. So you can spend more time analyzing the time series and less time training mundane Time Series forecasting models for your use case.

Wish to create one? Head on over to Tiyaro!

Share this