Introduction
For about 5 years, I've been working on a project fuelled only by my own passion and curiosity. A project built with the intent to save lives and reduce economic downfall.
This is a project that many have criticized and claimed to be impossible, but I had always believed there may just be a way of accomplishing, no matter how small the chances.
Throughout this project, I've faced many hardships regarding the challenges I have faced. Having to be self-critical to the highest degree throughout, constantly being in disbelief from my results. Being presented false-positive results, pseudo theories and critical mistakes have all been felt by myself.
We often see what we want to see, regardless of what the numbers may say. I've had to shut that out, look at only real data and graphs that do all the analysis for me. Approaching each branch of research with scepticism, thinking it likely won't work out.
I finally have what is close to a final answer, and I want to share it with the World. Welcome to the update on my earthquake forecasting project. Far from done, but showing even greater results than before.
What's New
I've continued to peruse the various leads I've had, many have fell off, some have made little to no difference even with VERY extensive testing and some have yielded some very interesting results. I've attempted to prove what is already known about earthquakes as I've went along, to ensure my data and hypothesis remains intact.
My recent development has focused around entropy to a more precise level, it has always been a core area of interest for this project; but it was like putting the cart before the horse. I stopped investigating so much until I had a more realistic model.
Therefore, I developed a completely new tool for this exact purpose. It could really be stated as a separate research project in of itself.: https://github.com/alexanderwalford-official/Dataset-Entropy-Tool
With varying levels of success using different methods and parameters, we observe surprisingly consistent results.
It may seem like I am going off-track here, but this relates directly to my earthquake research. The graphs you are about to see have never been released anywhere else before. All is exclusive to this specific blog post.
And of course, I'm going to keep things brief with images here. Naturally only showing you the results I am currently happy to share only.
Make what you will about the accuracy of these results, but more will be revealed soon. The proof of the success will be illustrated in the real forecasts.
I will soon be starting a new job, so my research may take a tumble yet again as I will have slightly less time to work on it. However, I plan to continue slowly working on this project until I have something I can definitively be confident in (mostly).
Next Plans
More testing of course! I will continue to analyse the entropy in my output dataset and attempt to lower the amount of entropy used in the inferencing method.
You will be the first to know about my next results, assuming I can be slightly more confident in them. I plan on eventually setting up a website and/or a YouTube livestream with real-time forecasts. The original plan was to develop a mobile application as well, but this is still very much hypothetical.
I will not share any of my "special numbers or techniques" at this current moment in time. The amount of time, effort and pain that I have poured into this project will keep me from releasing this any time soon. My proof will always be in the results, which should be enough to prove the accuracy to anyone.
This is an incredibly difficult project, I still do not know if it is 100% possible. It is VERY likely that my final forecasts will have a small amount of error margin but should still be quite useful. As far as I am aware, this is the first time anyone has used these specific techniques in this specific way to forecast earthquakes.
Conclusion
The reality of earthquakes is that they are truly random to a minor degree. Less than I had originally anticipated, but still more than enough to be a large problem.
Anything that my model will do, will always be an educated guess. There are also many other factors at play that I simply am incapable of observing at this current moment in time.
This does not mean that we can't accomplish our goal and provide a somewhat trust worthy forecast that could be used in concentrated areas across the globe. There have been many times I've doubted myself and my abilities, but now I'm feeling slightly more confident from my latest results.
I've still got a lot of work to do, this project feels like it keeps getting more and more complicated and keeps getting longer every time I look at it but I know that there should be an end in sight.
My final words for this post are: watch this space!
Comments
Post a Comment