This repository contains code and associated files for building and deploying machine learning and deep models using AWS SageMaker. This repository consists of a number of projects including tutorial notebooks. This repository has been developed as part of Machine Learning Engineer udacity Nanodegree.
-
Building a Model using XGBoost in Amazon's SageMaker This module contains a project to build a sentiment analysis model using xgboosta in SageMaker. It also contains tutorials explaining how to build xgboost models to predict Boston housing prices using Amazon's SageMaker service and SageMaker's built-in XGBoost algorithm.
-
Deploying a Model to Production in Amazon's SageMaker This module contains a project to deploy a XGBoost model and Sentiment analysis model using Amazon's SageMaker and use it with a simple web application. It also contains tutorials explaining how to build and deploy an xgboost model to predict the median value of a home in the area of Boston Mass.
-
Tuning Hyperparameter with Amazon's SageMaker This module contains a project to automatically tune XGBoost hyperparameter using Amazon's SageMaker to predict the sentiment of a movie. It also contains tutorials explaining how to tune hyperparameter of xgboost model to predict Boston housing prices using SageMaker.
-
Updating a Model in production with Amazon's SageMaker This module contains a project to update XGBoost model to predict the sentiment of a movie. The update is to meet changes in the underlying data used to train the original model. It also contains tutorials explaining how to update models in SageMaker without interrupting service.
-
Creating a Sentiment Analysis Web App using PyTorch and SageMaker This module contains a project to create a web app performing sentiment analysis on movie reviews. It contains preparing and transforming Data using NLP concepts, train a LSTM neural network using PyTorch in SageMaker and deploy the model for the web app. The project also uses AWS services such as Lambda, IAM roles and API Gateway to access SageMaker endpoint.
-
Population Segmentation This module contains a project to train and deploy models using PCA and k-means clustering to group US counties by similarities and differences. It contains visualizing of trained model attributes and interpreting the results.
-
Payment Fraud Detection This module contains a project to build and deploy a supervised, LinearLearner model in SageMaker. The project shows how to tune a model and handle a case of class imbalance to train a model to detect cases of credit card fraud.
-
Deploying PyTorch Neural Network This module contains a project to train and deploy a custom PyTorch neural network that classifies "moon" data; binary data distributed in moon-like shapes.
-
Time-Series Forecasting This module contains a project to analyze time series data and format it for training a DeepAR algorithm; a forecasting algorithm that utilizes a recurrent neural network. Train a model to predict household energy consumption patterns and evaluate the results.
-
Plagiarism Detector This module contains a project to build an end-to-end plagiarism classification model. It has part for cleaning data, extracting meaningful features, and deploying a plagiarism classifier in SageMaker.