So why do this at all? One of the problems is we face is the uncertainty of events in the future. Another problem we face is learning from events, and how we can better understand events, and the things that contribute to them. Another problem is how the amount of data we can collect. Plus how much data is specific to an event, rather than getting data that describes the circumstances/factors that caused the event to occur. I believe that there is just too many limiting factors when you consider the individual. But what if we reverse the process. What if we decide to investigate the circumstances that allowed / cause and effect / contributed the event to happen in the first place. If no event can occur with out the right circumstances, then surely those conditions are the best options to understanding the event. Another point is that there are less limitations, and more data available to describe the circumstances than there is in describing the individual.
One of the things I keep hearing about when I profile, is that Artificial Intelligence is the way of the future in prediction. Artificial Intelligence can tell you when some one is most likely to sit on a chair, however it will always struggle under current capabilities about who will sit on the chair.
Lets look at some common events that we collect data on,
breaking the law,
who have a car accident,
chatting someone up at a bar,
or catch covid or the common cold,
their house catches fire,
what they buy when shopping,
etc.
Let’s say for a point of discussion that we collect data on 100 people in general. In our 100 people sample population, there are people who have experienced at least one but maybe more of our scenarios. If one of the people sampled have experienced one of the scenarios we collect data about these events and the decisions made which contributed to these events. But we only directly gather data on their involvement in these events. What we don’t capture is all the data that contributed to the event. We also do not capture data about when the event didn’t happen.
Let’s say for example that of our 100 people, 10 of them purposely drove through a red light, and 5 of them accidently drove though the red light. What we don’t record is any data on the 85 people who didn’t drive through the red light.
For an Artificial Intelligence to be trained properly, it needs to know data on these three questions.
- Why did the person choose to drive though the red traffic light.
- Why did the person choose not to drive through the red traffic light.
- Why did the person not have to choose between either driving or not driving through the red traffic light.
The big issue with trying an analyse any event, is getting enough data on what was involved in the event. But we forget or think it is too impossible to try and include data that surrounds the event, but isn’t directly part of it. Part of the reason is the difficulty in getting data in the first place. But if we revisit the concept, that nothing happens without the opportunity for it to happen. Perhaps we should primarily be focusing not on the individual aspects but on the environment aspects first and then secondarily focus on the individual.
