>>9878955>IF IT IS ONLY ACCURATE 80% OF THE TIMEThis is easy. If you predict whether we are at war or at peace using a coin flip, you would be right 50% of the time (i.e. we've been at war for half our history). We could improve up to 80% by conditioning our prediction on variables like oil prices, troop movements. Another approach could be training a neural net to make predictions using news stories.
However, i think your wasting your time. Your models wouldn't tell you anything that wasn't obvious. You're probably better off reading the stories yourself. When the model tells us the war will happen, we already know
What you want is model tells us that a war will happen before we have a sense its coming. This would require a mathematical theory predicting when people want to start wars. Personally, I think there is too much 'free-will' to construct such a model.
t. economist (in training) i'm okay with error. In fact, its what makes economics exciting
