In the recent times technology is not limited to a vendor or product, open source is one arena where data mining and data warehousing techniques opened up new ventures and methods to look for patterns than formulated results. This is where Machine Learning (ML) elevates its importance in defining algorithms and system methods from the processing data to analyse. The bottomline is more data is better managed on algorithms/formulas, hence the data science rotates around this subject to design algorithms and computational formulas. One best example to identify how ML is useful, Google’s self-drive cars, speech recognition in devices (iOS Siri, Hello Google and MicrosoftCortana) and device intellgence methods (Google’s Project Tango) and so on.
There is no doubt that majority of mankind is fully dependant on gadgets/machines, so think about how it can be beneficial for business that will turn into data sets to evaluate and improvement for new methods. Think about consumer use of Twitter/Facebook/Whatsapp/Instagram applications and their feeds generating volume of data, where it helps developers to design the pattern according to your usage & behaviour (likes and dislikes) from those applications.
No doubt that Machine Learning is vast arena of study & opportunity, just to open the bonnet the key Machine Learning areas are differentiated into 4: supervised, unsupervised, reinforcement and deep. Further read on these areas from How Machine Learning Could Result In Great Applications for Your Business blog.
The key finding is analysis of data sets, by looking at number of variables within the complex data with the help of technologies such as Big Data distribution. Unlike in traditional Business Intelligence & Data Warehousing world where sample data or conceptual data is necessary to run certain models, there is no necessity with Big Data that you need/look for sample sizes and with the help of Hadoop distribution it is posisble to run through Internet of Things (IoT) (interesting read about IoT) which lends us Reactive Analytics, Proactive Analytics and Predictive Analytics. This will help us to find/explore hidden areas of analytics from that data set without having any issues on performance aspects.
Any bigger idea will always start small, so its better to think big and start small! Here are few links I would like to share with you to begin:
- Microsoft Research on Machine Learning projects.
- Introduction to Bigdata/Machine Learning
- Unsupervised learning – Google’s neural network
- Showcase of Success with Reinforcement learning
- Gartner’s Hype Cycle for emerging technologies
- Microsoft’s IoT solution – Azure IoT Suite & Microsoft’s IoT
This is just a beginning, there is more to learn…..