{"id":1930,"date":"2022-03-04T16:43:14","date_gmt":"2022-03-04T16:43:14","guid":{"rendered":"https:\/\/salarydistribution.com\/machine-learning\/2022\/03\/04\/build-a-cold-start-time-series-forecasting-engine-using-autogluon\/"},"modified":"2022-03-04T16:43:14","modified_gmt":"2022-03-04T16:43:14","slug":"build-a-cold-start-time-series-forecasting-engine-using-autogluon","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2022\/03\/04\/build-a-cold-start-time-series-forecasting-engine-using-autogluon\/","title":{"rendered":"Build a cold start time series forecasting engine using AutoGluon"},"content":{"rendered":"
\n

Whether you\u2019re allocating resources more efficiently for web traffic, forecasting patient demand for staffing needs, or anticipating sales of a company\u2019s products, forecasting is an essential tool across many businesses. One particular use case, known as cold start forecasting<\/em>, builds forecasts for a time series that has little or no existing historical data, such as a new product that just entered the market in the retail industry. Traditional time series forecasting methods such as autoregressive integrated moving average (ARIMA) or exponential smoothing (ES) rely heavily on historical time series of each individual product, and therefore aren\u2019t effective for cold start forecasting.<\/p>\n

In this post, we demonstrate how to build a cold start forecasting engine using AutoGluon AutoML for time series forecasting<\/a>, an open-source Python package to automate machine learning (ML) on image, text, tabular, and time series data. AutoGluon provides an end-to-end automated machine learning (AutoML) pipeline for beginners to experienced ML developers, making it the most accurate and easy-to-use fully automated solution. We use the free Amazon SageMaker Studio Lab<\/a> service for this demonstration.<\/p>\n

Introduction to AutoGluon time series<\/h2>\n

AutoGluon<\/a> is a leading open-source library for AutoML for text, image, and tabular data, allowing you to produce highly accurate models from raw data with just one line of code. Recently, the team has been working to extend these capabilities to time series data, and has developed an automated forecasting module that is publicly available on GitHub<\/a>. The autogluon.forecasting<\/code> module automatically processes raw time series data into the appropriate format, and then trains and tunes various state-of-the-art deep learning models to produce accurate forecasts. In this post, we demonstrate how to use autogluon.forecasting<\/code> and apply it to cold start forecasting tasks.<\/p>\n

Solution overview<\/h2>\n

Because AutoGluon is an open-source Python package, you can implement this solution locally<\/a> on your laptop or on Amazon SageMaker Studio Lab. We walk through the following steps:<\/p>\n

    \n
  1. Set up AutoGluon for Amazon SageMaker Studio Lab.<\/li>\n
  2. Prepare the dataset.<\/li>\n
  3. Define training parameters using AutoGluon.<\/li>\n
  4. Train a cold start forecasting engine for time series forecasting.<\/li>\n
  5. Visualize cold start forecasting predictions.<\/li>\n<\/ol>\n

    The key assumption of cold start forecasting is that items with similar characteristics should have similar time series trajectories, which is what allows cold start forecasting to make predictions on items without historical data, as illustrated in the following figure.
    \"\"<\/a><\/p>\n

    In our walkthrough, we use a synthetic dataset based on electricity consumption, which consists of the hourly time series for 370 items, each with an item_id<\/code> from 0\u2013369. Within this synthetic dataset, each item_id<\/code> is also associated with a static feature (a feature that doesn\u2019t change over time). We train a DeepAR<\/a> model using AutoGluon to learn the typical behavior of similar items, and transfer such behavior to make predictions on new items (item_id<\/code> 370\u2013373) that don\u2019t have historical time series data. Although we\u2019re demonstrating the cold start forecasting approach with only one static feature, in practice, having informative and high-quality static features is the key for a good cold start forecast.<\/p>\n

    The following diagram provides a high-level overview of our solution. The open-source code is available on the GitHub repo<\/a>.
    \"\"<\/a><\/p>\n

    Prerequisites<\/h2>\n

    For this walkthrough, you should have the following prerequisites:<\/p>\n

    Log in to your Amazon SageMaker Studio Lab account and set up the environment using the terminal:<\/p>\n

    \n
    cd sagemaker-studiolab-notebooks\/ \ngit clone https:\/\/github.com\/whosivan\/amazon-sagemaker-studio-lab-cold-start-forecasting-using-autogluon\nconda env create -f autogluon.yml\nconda activate autogluon\ngit clone https:\/\/github.com\/yx1215\/autogluon.git\ncd autogluon\/\ngit checkout --track origin\/add_forecasting_predictor<\/code><\/pre>\n<\/p><\/div>\n
    These instructions should also work from your laptop if you don\u2019t have access to Amazon SageMaker Studio Lab (we recommend installing Anaconda on your laptop first).<\/p>\n
    When you have the virtual environment fully set up, launch the notebook AutoGluon-cold-start-demo.ipynb<\/code> and select the custom environment .conda-autogluon:Python<\/code> kernel.<\/p>\n
    Prepare the target time series and item meta dataset<\/h2>\n
    Download the following datasets to your notebook instance if they\u2019re not included, and save them under the directory data\/<\/code>. You can find these datasets on our GitHub repo<\/a>:<\/p>\n