How custom models work - Amazon SageMaker AI

How custom models work

Use Amazon SageMaker Canvas to build a custom model on the dataset that you've imported. Use the model that you've built to make predictions on new data. SageMaker Canvas uses the information in the dataset to build up to 250 models and choose the one that performs the best.

When you begin building a model, Canvas automatically recommends one or more model types. Model types fall into one of the following categories:

  • Numeric prediction – This is known as regression in machine learning. Use the numeric prediction model type when you want to make predictions for numeric data. For example, you might want to predict the price of houses based on features such as the house’s square footage.

  • Categorical prediction – This is known as classification in machine learning. When you want to categorize data into groups, use the categorical prediction model types:

    • 2 category prediction – Use the 2 category prediction model type (also known as binary classification in machine learning) when you have two categories that you want to predict for your data. For example, you might want to determine whether a customer is likely to churn.

    • 3+ category prediction – Use the 3+ category prediction model type (also known as multi-class classification in machine learning) when you have three or more categories that you want to predict for your data. For example, you might want to predict a customer's loan status based on features such as previous payments.

  • Time series forecasting – Use time series forecasts when you want to make predictions over a period of time. For example, you might want to predict the number of items you’ll sell in the next quarter. For information about time series forecasts, see Time Series Forecasts in Amazon SageMaker Canvas.

  • Image prediction – Use the single-label image prediction model type (also known as single-label image classification in machine learning) when you want to assign labels to images. For example, you might want to classify different types of manufacturing defects in images of your product.

  • Text prediction – Use the multi-category text prediction model type (also known as multi-class text classification in machine learning) when you want to assign labels to passages of text. For example, you might have a dataset of customer reviews for a product, and you want to determine whether customers liked or disliked the product. You might have your model predict whether a given passage of text is Positive, Negative, or Neutral.

For a table of the supported input data types for each model type, see Custom models.

For each tabular data model that you build (which includes numeric, categorical, time series forecasting, and text prediction models), you choose the Target column. The Target column is the column that contains the information that you want to predict. For example, if you're building a model to predict whether people have cancelled their subscriptions, the Target column contains data points that are either a yes or a no about someone's cancellation status.

For image prediction models, you build the model with a dataset of images that have been assigned labels. For the unlabeled images that you provide, the model predicts a label. For example, if you’re building a model to predict whether an image is a cat or a dog, you provide images labeled as cats or dogs when building the model. Then, the model can accept unlabeled images and predict them as either cats or dogs.

What happens when you build a model

To build your model, you can choose either a Quick build or a Standard build. The Quick build has a shorter build time, but the Standard build generally has a higher accuracy.

For tabular and time series forecasting models, Canvas uses downsampling to reduce the size of datasets larger than 5 GB or 30 GB, respectively. Canvas downsamples with the stratified sampling method. The table below lists the size of the downsample by model type. To control the sampling process, you can use Data Wrangler in Canvas to sample using your preferred sampling technique. For time series data, you can resample to aggregate data points. For more information about sampling, see Sampling. For more information about resampling time series data, see Resample Time Series Data.

If you choose to do a Quick build on a dataset with more than 50,000 rows, then Canvas samples your data down to 50,000 rows for a shorter model training time.

The following table summarizes key characteristics of the model building process, including average build times for each model and build type, the size of the downsample when building models with large datasets, and the minimum and maximum number of data points you should have for each build type.

Limit Numeric and categorical prediction Time series forecasting Image prediction Text prediction

Quick build time

2‐20 minutes

2‐20 minutes

15‐30 minutes

15‐30 minutes

Standard build time

2‐4 hours

2‐4 hours

2‐5 hours

2‐5 hours

Downsample size (the reduced size of a large dataset after Canvas downsamples)

5 GB

30 GB

N/A

N/A

Minimum number of entries (rows) for Quick builds

2 category: 500 rows

3+ category, numeric, time series: N/A

N/A

N/A

N/A

Minimum number of entries (rows, images, or documents) for Standard builds

250

50

50

N/A

Maximum number of entries (rows, images, or documents) for Quick builds

N/A

N/A

5000

7500

Maximum number of entries (rows, images, or documents) for Standard builds

N/A

150,000

180,000

N/A

Maximum number of columns

1,000

1,000

N/A

N/A

Canvas predicts values by using the information in the rest of the dataset, depending on the model type:

  • For categorical prediction, Canvas puts each row into one of the categories listed in the Target column.

  • For numeric prediction, Canvas uses the information in the dataset to predict the numeric values in the Target column.

  • For time series forecasting, Canvas uses historical data to predict values for the Target column in the future.

  • For image prediction, Canvas uses images that have been assigned labels to predict labels for unlabeled images.

  • For text prediction, Canvas analyzes text data that has been assigned labels to predict labels for passages of unlabeled text.

Additional features to help you build your model

Before building your model, you can use Data Wrangler in Canvas to prepare your data using 300+ built-in transforms and operators. Data Wrangler supports transforms for both tabular and image datasets. Additionally, you can connect to data sources outside of Canvas, create jobs to apply transforms to your entire dataset, and export your fully prepared and cleaned data for use in ML workflows outside of Canvas. For more information, see Data preparation.

To see visualizations and analytics to explore your data and determine which features to include in your model, you can use Data Wrangler’s built-in analyses. You can also access a Data Quality and Insights Report that highlights potential issues with your dataset and provides recommendations for how to fix them. For more information, see Perform exploratory data analysis (EDA).

In addition to the more advanced data preparation and exploration functionality provided through Data Wrangler, Canvas provides some basic features that you can use:

  • To filter your data and access a set of basic data transforms, see Prepare data for model building.

  • To access simple visualizations and analytics for feature exploration, see Data exploration and analysis.

  • To learn more about additional features such as previewing your model, validating your dataset, and changing the size of the random sample used to build your model, see Preview your model.

For tabular datasets with multiple columns (such as datasets for building categorical, numeric, or time series forecasting model types), you might have rows with missing data points. While Canvas builds the model, it automatically adds missing values. Canvas uses the values in your dataset to perform a mathematical approximation for the missing values. For the highest model accuracy, we recommend adding in the missing data if you can find it. Note that the missing data feature is not supported for text prediction or image prediction models.

Get started

To get started with building a custom model, see Build a model and follow the procedure for the type of model that you want to build.