Loud ML is the first open source deep learning API that makes it simple to prepare, train, and deploy machine learning models and crunch the data stored in your favorite databases without moving the data. The user selects the times series that they want to model and sets the model date ranges, then Loud ML will build the models and save them for inference in production. Loud ML does all the work and removes the complexity of machine learning with Tensorflow.
Here are a few sample use-cases that Loud ML is used for:
- Detecting abnormal dips in user traffic and responding to incidents before they impact customers satisfaction
- Detecting outliers in seasonal fluctuations of e-commerce transactions
- Spotting abnormal load in a distributed database
- Dynamically spotting network traffic patterns and anticipating congestion before it impacts customer experience
- Forecasting capacity, usage, and load imbalance for energy producers and suppliers
- Forecasting demand for inventory and supply chain optimization
- Abnormal fraud pattern detection for mobile network operators
- Predict network equipment failure for maintenance operations planning
- Anticipate disk capacity and discover capacity issues before it hurts
- Knowing the future load in advance and auto scaling virtual instances in the Cloud
- Knowing the future load in advance and saving energy in data centers
Loud ML ships with
unsupervised learning techniques that do not require labelled data and therefore can produce faster results.
Donut [arXiv 1802.03903](https://arxiv.org/abs/1802.03903) combines the best of unsupervised and supervised learning: users can label abnormal data if they want to, although this label operation remains optional.
For the rest of this tutorial, you will be guided through the process of getting Loud ML up and running, taking a peek inside it, and performing basic operations like creating, training, and using your data to get accurate predictions. At the end of this tutorial, you should have a good idea of what Loud ML is, how it works, and hopefully be inspired to see how you can apply ML to your own data and application.