Add to Watchlist

Vector Transmissions of Disease with Nupic [DEMO #10]


Citation of segment
Embed Code
Purchasing a DVD Cite video

Formal Metadata

Title Vector Transmissions of Disease with Nupic [DEMO #10]
Title of Series 2015 Spring NuPIC Hackathon
Number of Parts 19
Author Winget, Jean
Phuvan, Vietnhi
License CC Attribution 3.0 Unported:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
DOI 10.5446/18067
Publisher Numenta Platform for Intelligent Computing (NuPIC)
Release Date 2015
Language English

Content Metadata

Subject Area Computer Science
Abstract Trying to use NuPIC to understand the transmission of Ebola.
all of them in the and so on and how
this could we wanted to do it there's a book with some disease vectors that I have worked on earlier this year for the variable a project that on we're analyzing data and trying to basically located you know how to describe this a couple of factors 1 was trying to normalize the data through many many many sources and many variations in accuracy levels and and rates of return and the other was simply trying to then read you know solution and graphic data so obvious you know this is the data that came out at that time and you can see that there several anomalies in the construction of the data so talking with everyone here we thought it might be pretty interesting to read plot this data using the hot yeah structure to see if we can actually you know physically see the anomalies in the new mixed system how's it going to move I have to number of screen that and 2 of his sense of what was going so this is another plot that we did you know earlier in the initial phase of the data analysis of back with this is so this is our final plot on on and needlessly the technical details on the Python and and the additions to the to the actual data that we ran through and explain where the data how we do the analysis on the data but this is basically as you can see once again we're capturing the anomalies in the data you can see if you wanna move back again you can see the areas where we have a drop out of actual physical 1 of the problems is if you notice in this region right here at consider what life is on you have actual physical dropouts where no data for whatever those
transmissions you know we're dealing with countries that they have you know very difficult transmission of satellite things happen right is not like
America that is how it it in the it was difficult to try to decipher and try to encode trying this is basically what I was trying to do is try to find a normalized regime so that maybe we could use this information to get more accurate 0 this is a
mathematical models those working with and others with I I work concerns and as as the model which is actually kind of closely representative to stock model in the original because the idea that I talked was I was like OK well hurricane capture this you know data is it's so erratic but we have we have to have some kind of common factor here and that the contractor was much like stock data in the there's is going up or it's going down will unfortunately
here basically the initial premises theories are alive or you know what the hell on inside of that data there's about 5 other tertiary steps of the probability diagnosis that it was in fact bowl of over on this side of a number of other options so to speak of how when people they have a 21 day window that we were trying to capture also from the time that they possibly were infected and then there's the actual outcome of people of cases that were in fact infected and then there's more about you know the mortality in the actual recovery rate also so there's a huge capacity equations you use to calculate from the status random takes this capacitive differentials from the normalized data from this log log plots so anyway so we got back here and this is what we're showing is that we're definitely there system is definitely capturing the same drop-off points and the same anomalies so it looks like between the 2 plotting systems between the mathematical planning system and they're plotting system their probability matrix we're getting very close similar relationships which is the good grades in the same so the next thing that I would really like to be able to start working on and we just basically
discusses is working on the actual cell data because 1 of the things that's happening is we're getting many many data points on 1 day from various areas so we like to start training and tracking the actual physical movement of the disease so that we would know 0 is probably going to x next is probably going here's how we going there the probability of a going here is what's likely demand so this is the just the initial I just started working on very simple matrix algorithm plotting what could possibly turn into a neuron cell relationships such physically everything we got due you unintelligible there are hormones and they to the the it was a the little beginning it was based on the way how can I do is I mean here you can use the same form just 1 yeah it was found that there we go say
so that's it thank you very much for your mind


  335 ms - page object


AV-Portal 3.8.0 (dec2fe8b0ce2e718d55d6f23ab68f0b2424a1f3f)