2 Likes
Oh Peter, would you please explain that thingy in more laymanâs term for us mere mortals. 
To think of Peter Griffin explaining machine learning is really entertaining. That said, as someone who only really picked up programming in the last year (and previously thought âthat looks like too much workâ) I can definitely empathize with the âwhat the hell is this?â perspective.
What the hell is Python? Ironically, in the programming world, Python is actually quite analogous to the role that Gini played â it is useful in a bunch of different areasâŠfrom making GUIs (applications) to web development and pretty much everything in between. Particularly, it is useful with machine learning. Also, from the text of the code, it is the closest thing to English that you will see in programming. R, another language that I use, looks basically like code vomit, whereas Python is (comparatively) cleaner.
So why the hell did you post this? There is a Python model called RandomForestClassifier (RFC), which is essentially a model that looks at labeled outcomes (e.g. âBLOWOUT RESULTâ, âCLOSE RESULTâ, âDRAWâ), looks at all of the information that determines those outcomes (e.g. goals), and then predicts future outcomes based on what it learns. Itâs pretty cool! In fact, I taught myself programming by using the FiveThirtyEight soccer datasets, and just trying new concepts as I learned more.
You still didnât explain the gini bit: Anyway, in programming, models have settings that you can configure. One of these settings is called âcriterionâ, which means âhow do you want the model to improve itself?â You can enter âentropyâ or âginiâ, the latter of which essentially means âcalculate the probability that the machine is wrong when it makes a random pick.â The idea is for the machine to calculate âginiâ and then use it to make better predictions by figuring out what leads to wrong predictions. I have set RFCs to have âginiâ most of the time, but â instead of entering âginiâ â I tied the string âginiâ to the word wijnaldum. Did it purely as a joke, but I still got more entertainment out of it than I should have done.
Not sure if you meant that seriously, @Iftikhar , but perhaps you learned something today. 
2 Likes
I come here for football chat and even that confuses me sometimes.
4 Likes
At least we now know Cardiff Peteâs last name.
8 Likes
Well Iâm glad we cleared that up
1 Like
Yay, Monty Python.
Alright, Iâll try to get us back on track: El-Hadji Diouf
1 Like
Pepe Reina on The Masked Singer.
5 Likes
Suppose he must be doing it for a laugh, canât need the money i would hope.
2 Likes
To play him at LB.
He needed the costume so he could copy Sterlingâs runningâŠ
I wonât lie, the battery situation aside I like Flanno. Scouse Cafu, if you will.
long live his tackle on Soldado.
1 Like
is that possible though?
its one of those great philosophical questions.
1 Like
no itâs not. lost a lot of respect for him and it really was the demise of his career
2 Likes
At 39 he stills at some professional football
Always liked him, others scored more goals but he gave us juale le vivre.
4 Likes
Joe Cole on his time at Liverpool FC: "It just didnât work and Iâm gutted mate.
"I would have loved it to work and I would have loved to have played with [Luis] Suarez and [Raheem] Sterling in a front three.
âIn my prime I could have really complimented them, I think we would have been good but my body just wouldnât let me do it.â #LFC
How many months have he been on his prime? Worst free signing we ever had in the last 20 years.
1 Like
About 2002-09, a period that eneded when he fucked his knee.
2 Likes
Joe Cole was class. Should have been good for us butâŠ. Thatâs football
5 Likes