I would like to apologize at once to those who spent years of researching AI areas, such as speech recognition, or working on AI in trading systems. Giving myself over to thoughts, I would like to start a subject about this.
In the scientific world, one starts out from seeking correct definitions of terms, which takes half of all the energy. The other half goes to convincing of one’s truth. Say, if you talk about Data Mining, but use the terms that belong to statistic analysis, or even worse – to AI – sure enough, you’ll meet with deserts.
I suggest not to start out with the terms, but with the real business problems, and afterwards, do classification by using terms.
So my advice is to start with a simple example:
Description of a problem
An owner of 10 hotels chain wants to get a bigger profit (I promised – it is a very simple example).
Option А – we will ask a beautiful girl (hotel manager) to do analysis
1. Collect information /*date, name, last name, payment, hotel, room, breakfast, bar, sightseeing, etc. */
2. Make analysis and find associations and matches
3. Improve something in our business model (for example, make breakfast cheaper, but bar drinks more expensive)
4. Wait for results and check the bank account
Option B – (AI)
In this case, we can make a decision and examine it in a more professional way, can’t we?
Thus, let’s move our AI to the more practical field – a help system in decision making for the manager.
Description of the problem
On the basis of the huge amount of data /*date, name, last name, payment, hotel, room, breakfast, bar, sightseeing, etc. */ it is difficult to understand where and what should be improved.
Solution А – (again we ask the girl – the hotel manager)
1. She asks a programmer to provide an excel sheet with everything in it (2 days)
2. Makes 10 analyses on the basis of her ideas and experience (3 days)
3. offers her solution to the executives (2 days to create and 3 days to explain)
Solution B – (AI)
Open the AI instruments panel, whereby simple manipulations we get prices, breakfasts, sightseeing, etc., i.e. the set of criteria possibly important in decision making.
1. Choose an option (room price), the tip pops up – to reduce, to increase, as well as the list of secondary factors
2. Choose the price of a Single Room and see the relative price of the competitors.
Based on the history we can assume what is going to happen after making one or another decision. I.e. by changing the current price of a single room the system can predict how it is going to affect the secondary parameters (for example, hotel occupancy in the nearest future, and, accordingly, the profit). Quite uncomplicated algorithms are used for that.
Everything is performed in a visual way, and, accordingly, is understandable both to the girl and the executives. Below is an example of such instruments panel.
3. On the basis of all this information we prepare an offer to the executives, which looks like this:
The price of a Single Room is 87 USD, breakfasts are 10 USD. (We can add the competitor’s data, attach a report with a diagram). We are planning to have this and that data in the nearest months.
As you can see, the system helps to ease decision making, as it contains the set of data in a convenient form. The question remains only about to have the opportunity to collect and store all the necessary information, as well as provide a convenient solution to display this information. The one and the others are relatively easy to describe and develop.
At this point, I would like to ask you a question. If I am not mistaken, are we talking here about Data Mining methodology to collect data, statistical analysis algorithms to analyze past data and Artificial Intelligence methods to make prognoses for future?
I would be very grateful if you could help to expand this example and review the use of AI in our world by using obvious and simple examples.