By Aldis Ērglis, Intelligent Machines Riga

 

We are already experiencing an impact of Machine Learning approach in business. We see it in Google search, on Amazon or Netflix recommendations. It means companies with large data amounts can feed learning algorithms with right data and get amazing results in decision making or decision support field.

 

Machines will learn faster and more accurately than humans for sure, and we are not talking about smart Robocop, AI like machines but even simple statistical algorithms, like described in Daniel Kahneman, HBR article Noise – https://hbr.org/2016/10/noise. Even more, most of algorithms and models are available and most are for free, like more than 12000 R packages available on the internet.

 

So, technically all companies can start using Machine Learning to create additional products and services, add intelligence to existing products and services or reduce costs by automation even in decision-making field but mostly in decision preparation for decision makers.

 

Why cannot all companies do that? Because Machine Learning needs a lot of good quality data to learn right things. Few companies are already collecting or let say mining data from their processes, customers, employees, and systems for Machine Learning purpose. And believe or now for good ML different data is needed. For example, to analyse and learn from processes you need the collect duration of operations instead of date and time of state. That is different data than ERP currently collecting because the purpose of data is different. So, it means to implement Machine Learning we need to rethink all data we are collecting and start collect data for ML purpose. Sources of data will be behind your organization boundaries, additionally to data inside the organization you need to collect data from the internet about your customer – using Competitive Intelligence approach and so on.

 

Is data a new Oil? Forbes argues in the recent article (https://www.forbes.com/sites/bernardmarr/2018/03/05/heres-why-data-is-not-the-new-oil/#1af3b1973aa9) that data is not new Oil. Maybe comparing very precisely it is not. But I see some analogies that will help business leaders to think about data in the better way. To run the economy as we use a lot of raw materials, oil is needed, and during years we found how to discover, get, and use oil. The same idea with data, you can’t use any data you have, you need to discover, build up mining process, control quality, distribute it and use for the right purpose. Not all oil we can get from earth is usable so the data even more.

 

I see that organizations will be under pressure to change their mind because of competitive advantage of using ML. So, companies need to start with thinking how ML can help achieve business strategy and create ML strategy. ML strategy will show in what areas, processes you need to collect good data. And even if you are not doing any ML, find out and start collecting data, it will take time and will not be possible to do it overnight. So, collecting right data is your competitive advantage.

 

Visit the RIGA COMM 2018 Machine Learning Practical Application Conference on 11 October to find out more.

Follow the conference page: https://www.facebook.com/events/432218577193046/