The purpose of machine learning is often to predict or anticipate something. Machine learning can, for example, be used to predict future developments, actions or results with a far greater level of probability than we as people can work out ourselves. The technology can also be used for facial, image, text and speech recognition.
The data that the computer depends on can, for example, come from public sources, server activity, IoT devices, customer activity and social networks. The amount and quality of data influence how quickly and how successfully the program learns.
Machine learning has enormous commercial potential in that it enables a range of opportunities to increase revenue and reduce costs. Some example applications of machine learning include:
1. Identifying fraud and predicting risk in connection with loans and insurance. In combination with data streaming, entirely new data can be included and it will be possible for predictions to be made in real-time.
2. Anticipating or predicting the results of elections, the direction of equity markets and oil prices, influenza pandemics, pollen concentrations, and the weather.
3. Anticipating which customers are the most open to buying more products, or which will probably leave us if we do not do something to stop this.
4. Communicating in chatbots.
5. Making mobile payments seamless with the help of image recognition.
6. Making it easier to search for and purchase items online, and making our homes smarter, using speech recognition.
7. Providing more relevant search results online.
8. Identifying false or hate-speech posts or undesirable images, e.g. as Facebook does.
9. Finding the best route when there are numerous possible routes to a destination, e.g. on the basis of distance, speed limits and traffic jams. Google Maps, for example, uses machine learning in combination with data streaming.
10. Identifying errors in data and internal processes.
A number of cloud services have resources that now make it easier to get started with both machine learning and advanced data analytics. These two technologies are collectively referred to as data science.