Join Jonathan Fernandes for an in-depth discussion in this video What you should know before watching this course, part of AWS Machine Learning by Example.
- [Instructor] For this course, I would recommend you have a basic understanding of Machine Learning, but you won't need any prior experience with AWS. I also suggest that you get an AWS account if you don't have one as it's the most affective way to learn. And then you could try out the exercises for yourself. I'll show you how to set up an account in the next video. Alternatively, you're welcome to just follow along with the videos. AWS supports Machine Learning and real-time prediction endpoints in two regions, the US and Ireland.
That doesn't mean that you can't store your data sets in other regions. However, I would suggest you stick with one of these two regions for this course. There are two reasons for this. If when you run some of the real-time prediction endpoints from any region by calling an endpoint from a region that does not have that endpoint can have an impact on real-time prediction latencies. The other factor is that if you move data across regions, you will start to have cross region data transfer fees. Amazon Machine Learning charges an hourly rate for the compute time used to compute data statistics in train and evalued models.
And then you have to pay for the number of predictions generated for your application. For real-time predictions, you'll also pay an hourly reserved capacity charge based on the size of the model. So remember that there are no minimum fees and no upfront commitments so you only pay for what you use. Remember that when you store the data sets in Amazon Storage, which is S3, you will be charged a small fee. The way I have designed this course is to try and minimize the amount of AWS fees you'll have to pay. When we perform batch processing, you will normally have to pay 10 cents for every 1,000 transactions, and we will ensure that the data sets are small.
After we've set up a model and generated some predictions, I'll show you how you can clean up your AWS environment so that there are no further charges. There are a couple of ways to access Amazon's Machine Learning. You can access the Amazon Machine Learning Console by signing into the AWS Console, you can install and configure the AWS CLI, or you can access the Amazon Machine Learning API. I've designed this course with the Machine Learning Practitioner's view in mind. So we will use the AWS Machine Learning Console exclusively.
And we'll be more interested in analyzing, evaluating, and interpreting the models rather than using APIs or the CLI to access the Machine Learning's services on AWS. In the next video, I'll show you how to set up an AWS account if you don't have one.
- Learning algorithms and hyperparameters
- Preparing data for AWS
- Using binary, multiclass, and regression techniques
- Creating a datasource
- Generating predictions
- Creating and interpreting batch predictions
- Additional AWS capabilities