|#How to use Machine learning skills for Internet of Things IoT:|
Machine Learning comprises algorithms which learn from data, make predictions based on their learning, and have the ability to improve their outcomes with experience. Due to the enormity of data involved with Machine Learning, various technologies and frameworks have been developed to address the same. Hadoop is an open-source framework targeted for commodity hardware to address big data scale.
The distributed design of the Hadoop framework makes it an excellent fit to crunch data and draw insights from it by unleashing Machine Learning algorithms on it.
So, the true value of IoT comes from ubiquitous sensors’ relaying of data in real-time, getting that data over to Hadoop clusters in a central processing unit, absorbing the same, and performing Machine Learning on data to draw insights; all at petabyte scale or more.
In reviewing the use cases and challenges from preceding sections, one thing is very clear. That is to do with quickness with which certain analytics must be performed. Imagine sending a critical alert late because computing could not be done any faster. Two key gaps here include absorbing incoming data at such a high rate reliably and in observing that Hadoop was not created for real time streaming data.
It was originally envisaged as a framework for batch processing. Innovators have responded to those challenges well. Let us review some of those technologies now.
SAP HANA with internet of things came into picture with real time processing of data compared to Hadoop which is only batch processing.
Iot Basics Part-1