All Collections
Building AI Models
Model Breakdown: Use Model
Model Breakdown: Use Model
O
Written by Obviously AI
Updated over a week ago

This article discusses the second tab on your model report after the Summary tab as mentioned in the previous article. This tab actually lets you use your trained model to make future predictions by clicking on the Launch Model button.

AutoML

The main sections on this tab are as follows: The Launch Model button opens on a new tab where you can directly put in values for each feature and get single predictions by clicking on the Predict button. You can also upload a CSV file with multiple rows (file size 25MB max) and get multiple predictions together as a downloadable CSV file.

In case you have a classification problem, the prediction probabilities for each class is displayed on the left. You can also check the Model Tech Specs here which is elaborately presented as Tech Specs tab on the platform. Here you can quickly review it without having to go into the platform again

In case you have a regression problem, the numeric value along with the uncertainty range is displayed on the left. For example, 7.92, uncertainty +- 1.499 means that the predicted value ranges between (7.92 +1.499) and (7.92 - 1.499). You can upload a CSV file for making batch predictions or make single predictions using the interface on the right.

Next is the REST API section that allows you to integrate your trained model and set up real time automations with our APIs.

  • Clicking on the Documentation opens our developers.obviously.ai page that details all our API endpoints with their working code

  • Clicking on the Video Tutorials you are directed to our No-Code AI University page, where you can enroll for free and see detailed videos on how our platform works

  • You’ll also find your personal API key and the Model ID corresponding to your model

  • Finally you’ll find the working API endpoint code for single prediction in Python, Javascript and Curl. You can directly copy the code and use it as it is

Time Series

For time series we have the Use Model tab that has the following sections in it. Model Summary that gives a quick overview of the model such as the best performing model name, the value of the metric skill score and the last timestamp in your data. This helps because for every forecast that you perform it will start from the next date from the last timestamp

Launch Model section allows you to actually use your trained model and opens on a new tab.

On the launch model page you’ll find the following: Make Predictions and Model Tech Specs

Forecast Period: Default value is 10. Here you can put the duration of your future predictions. For example if you uploaded a file with monthly data and chose the Data Level monthly, then you’ll see the forecast values for the next 10 months as shown on the right on the Predictions section. The predictions can be downloaded as a file using the download button

Predictions plot: This gives a visual representation of your predictions. On the horizontal axis, you see the future dates and on the vertical axis you see the predicted values. Additionally you’ll also find the icons for zooming in or out of the plot and also downloading it

You can also check the Model Tech Specs here which is elaborately presented as Tech Specs tab on the platform. Here you can quickly review it without having to go into the platform again

Next is the REST API section that allows you to integrate your trained model and set up real time automations with our APIs

  • Clicking on the Documentation opens our developers.obviously.ai page that details all our API endpoints with their working code

  • Clicking on the Video Tutorials you are directed to our No-Code AI University page, where you can enroll for free and see detailed videos on how our platform works

  • You’ll also find your personal API key and the Model ID corresponding to your model

  • Finally you’ll find the working API endpoint code for single prediction in Python, Javascript and Curl. You can directly copy the code and use it as it is.

To see Obviously AI in action, checkout this demo video OR enroll in the No-Code AI University for free to become a certified no-code AI expert.

Did this answer your question?