Algorithms

Examples on how to use SageMaker’s built-in algorithms.

Image processing

SageMaker provides algorithms that are used for image processing.

image_classification

• Using SageMaker Image Classification with Amazon Elastic Inference
• Image classification training with image format
• End-to-End Incremental Training Image Classification Example
• Image classification training with image format demo
• Image classification transfer learning demo
• Image classification transfer learning demo (SageMaker SDK)
• End-to-End Multiclass Image Classification Example
• End-to-End Multiclass Image Classification Example with SageMaker SDK and SageMaker Neo
• Image classification multi-label classification

object_detection

• Amazon SageMaker Object Detection for Bird Species

semantic_segmentation

• Amazon SageMaker Semantic Segmentation Algorithm

Text processing

SageMaker provides algorithms that are tailored to the analysis of texts and documents used in natural language processing and translation.

blazingtext

• Text Classification using SageMaker BlazingText
• Learning Word2Vec Subword Representations using BlazingText
• Learning Word2Vec Word Representations using BlazingText

lda

• An Introduction to SageMaker LDA

ntm

• Amazon SageMaker Neural Topic Model now supports auxiliary vocabulary channel, new topic evaluation metrics, and training subsampling
• Introduction to Basic Functionality of NTM
• An Introduction to SageMaker Neural Topic Model

seq2seq

• Machine Translation English-German Example Using SageMaker Seq2Seq

Time series processing

SageMaker DeepAR algorithm is useful for processing time series data.

deepar

• Time series forecasting with DeepAR - Synthetic data
• SageMaker/DeepAR demo on electricity dataset
• Predicting driving speed violations with the Amazon SageMaker DeepAR algorithm

Supervised learning algorithms

Amazon SageMaker provides several built-in general purpose algorithms that can be used for either classification or regression problems.

factorization_machines

• An Introduction to Factorization Machines with MNIST

knn

• Multi-Class Classification using Amazon SageMaker k-Nearest-Neighbors (kNN)

linear_learner

• An Introduction to Linear Learner with MNIST
• Train Linear Learner model using File System Data Source
• Build multiclass classifiers with Amazon SageMaker linear learner
• Fairness Linear Learner in SageMaker

xgboost

Basic

• Multiclass classification with Amazon SageMaker XGBoost algorithm
• Regression with Amazon SageMaker XGBoost algorithm
• Customer Churn Prediction with XGBoost

Advanced

• Train and deploy a regression model with the Amazon SageMaker XGBoost Algorithm using Script Mode
• Regression with Amazon SageMaker XGBoost (Parquet input)

Unsupervised learning algorithms

Amazon SageMaker provides several built-in algorithms that can be used for a variety of unsupervised learning tasks such as clustering, dimension reduction, pattern recognition, and anomaly detection.

ip_insights

• An Introduction to the Amazon SageMaker IP Insights Algorithm

kmeans

• Analyze US census data for population segmentation using Amazon SageMaker
• End-to-End Example with Amazon SageMaker K-Means
• End-to-End Example with Amazon SageMaker K-Means using SageMaker API

pca

• An Introduction to PCA with MNIST

rcf

• An Introduction to SageMaker Random Cut Forests

Feature engineering

object2vec

• Document Embedding with Amazon SageMaker Object2Vec
• An Introduction to SageMaker ObjectToVec model for MovieLens recommendation
• An Introduction to SageMaker ObjectToVec model for sequence-sequence embedding