Once stationarity was established, I proceeded to create a forecasting model for police shooting incidents. My model of choice was SARIMAX, which adeptly navigates the intricacies of seasonal and non-seasonal factors. Here’s an overview of my findings:
I relied on the AIC and BIC values as my guideposts for measuring the model’s effectiveness. Generally, the rule of thumb is the lower these numbers, the better. They indicated that my model struck the right balance: it was sufficiently sophisticated to discern the underlying trends and patterns, yet not excessively so as to overfit the statistical data.
However, the true essence of the story was revealed through the coefficients, especially those associated with the moving average components. The notable negative values of these coefficients suggested a compelling impact of recent occurrences on the probability of future events.
What I’ve crafted transcends a mere statistical model; it embodies a visionary lens into the timing and locations of police shootings. This extends beyond scholarly research; it touches upon the fabric of human lives, fostering the hope that through foresight, we might discover ways to avert such tragedies.