Bayesian belief networks

"Hey Zoe, I've heard you've started modelling"

One of my Cranfield colleagues said that to me the other day, and I'm fully aware I'm not camera shy, but if someone was using my face in some sort of modelling campaign then I wanted to know about it, so I could tell everyone! It was a few seconds into the conversation before I realised they were talking about my survey, modelling my data and Bayesian belief networks (BBN).

Now, I'm not an expert on BBN, as yet, so I have included a link to the wikipedia site which has some great links on it if you want to read a bit more into it.

What I know is that I can use BBN to model my data in a way that allows me to see the degree of belief that a child node will be in a particular state given the states of the parent nodes. This sentence took a while for me to understand, but if you see the picture below, you can see that the 'LocalIs' and 'TemporalScaleWork' boxes are the child nodes and the 'Policy' box is the parent node. 

If a respondent identified as working in Policy in the Private sector then it can be said that there is a 17% probability that they feel they work in the short term and an 83% probability that they work at the medium term, temporally. Their identification of what the term Local means is shown in the 'LocalIs' box. 

One of the reasons that I am using BBN is because of how interactive it is, by selecting the different options within the nodes you can show the interaction between the different bits of information. Another reason for using it is it is possible to feed in data from different sources, as it doesn't all need to be uniform data. This will be useful when I begin my workshops and want to integrate the data I collected.

I really like using BBN and I use the Norsys Netica software to build my networks, it's very user friendly, and has a great 'help' mode.One of the best things that has come of this so far is that I have managed to start answering my own questions about BBN, so I really feel like I'm getting a grasp of how it all works.

Achievements

Just having a think about all the little achievements that I have accomplished, as I am half way through my PhD, and struggling to stay sane. I'll go into more detail about each one, as and when they get even more exciting.

• I have an interview for a science policy internship in Westminster
• Just having the interview alone is so exciting, as I had to write a few different pieces of work for the application. Knowing that what I have written is good enough to get through to interview is a huge achievement.
• I have had my poster abstract accepted at the BESSfe conference in Lille.
• I get to go to Lille and use my poster as a catalyst for talking about my work, and I will hopefully make lots of really useful contacts.
• I have closed my survey
• No longer dedicating any time to pushing it, or finding new places to advertise it and try to make people do it. I really want this whole chapter of my thesis to be written, done and dusted by the end of this year. 
• I have (almost) mastered using Netica to build Bayesian Belief Networks.
• This is rather specific, and will mean something to some people, but nothing to others. I am rather proud of myself for this one as statistics have been the bane of my life since I began my academic career, so to actually know how to use a data interpretation programme is super exciting.
• I have made business cards
• When they come through I think that will be when I finally realise I am an adult.
• I am still half sane, and halfway through something that I constantly tell myself I can't do!