Originally posted by Hitchhiker's Guide To...View Post
afair ologit will tell you both overall significance and if the order is important - i.e. are 'more' false memories more related to explanatory variables.
mlogit will use a contrast value in dependent variable - probably '1', but you choose, and then whether the other two values are important individually and relatively to contrast value.
Might have misremembered the models a bit though.
When I read it first, I thought an ordinal logistic regression either probit or logit would suffice but then I thought I'm 100% missing something here because i never heard of a zero inflated negative binomial regression so I figured Hotspur definitely knew logit or probit.
If you end up using either, I believe little difference exists except for some variance in the tails
Test of parallel lines significant, proportional odds assumption rejected, a couple of variables the issue, could do partial proportional odds with gologit2 in Stata but don't use it, am fucked if I'm reducing it to multinominal regression, fuck it.
Test of parallel lines significant, proportional odds assumption rejected, a couple of variables the issue, could do partial proportional odds with gologit2 in Stata but don't use it, am fucked if I'm reducing it to multinominal regression, fuck it.
Steady decrease in the tone of hotspur's posts since his initial request.
Expecting some kind of LL-style mega rant by the end of day.
"We are not Europeans. Those people on the continent are freaks."
My commute is getting steadily worse. I'm leaving earlier than ever but I can't seem to beat the traffic. I left before 7am today and it took me 90 mins.
Combination of bad weather and N7 roadworks make it a nightmare.
If I present a representation of a fake news event to you and you say you remember seeing or hearing that then it's a false memory. It's actually an interesting area, especially in respect of false autobiographical memories where people are induced to falsely remember, say, a childhood hot air balloon ride.
The closest I got was Mandela Effect, then starting tying myself in knots if the the recollection of a false memory was more akin to suggestibility and the transference of another persons false memory being accepted as a memory . ie if we both share the same false memory is it your memory or mine?
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
Originally posted by Hitchhiker's Guide To...View Post
Zero-inflated negative binomial is actually a cool concept. Say you want to model a count of how many fish people who go to lake catch. The issue here is that your count data is capturing two types of people who go to a lake: those who go to a lake to catch fish, and those who just like going to lakes. So there will be two types of zeros in the model: from fishermen who caught no fish and those who didn't visit the lake to fish. Your zero counts in your model are then 'inflated'. So you nest a logit within a count model (negative binomial, poisson) and that models the zeros seperately, while still allowing them to play a role in estimation.
Quite useful in corporate finance - e.g. looking at acquisitions made by companies. Some will have tried to acquire but decided it wasn't a good idea, while others won't have even searched for acquisitions. So both those groups will be a zero count in your acquisitions model for different reasons.
speaking of fish....
Code:
A Population of Fish
There is a population of fish numbering 10000 at the start of the year.
Every month 5% of the fish alive at the start of the month, excluding offspring, die during the month.
For the months of May/June/July 25% of the fish alive at the start of the month pair off and pro-create. Each pair have 3 off-spring.
Every month, 15% of newly born fish over 3 months old depart the colony.
Assumptions:
Off spring do not pro-create
Make any other assumptions you need if you are in doubt.
Write code to calculate the number of fish at the end of the year
Every single person in here had a different answer
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
My commute is getting steadily worse. I'm leaving earlier than ever but I can't seem to beat the traffic. I left before 7am today and it took me 90 mins.
Combination of bad weather and N7 roadworks make it a nightmare.
I shudder at the thoughts of having to go near the N7.
I'm reasonably OK in that once I leave Ranelagh by 7:15, I'm in work at the back of the airport by 8. Going home is similar time frame but that's based on leaving work after 6.
I left work last Thursday at 4:30 to try get a few pints in before the Ireland game and it took 90 minutes to get back home!
My commute is getting steadily worse. I'm leaving earlier than ever but I can't seem to beat the traffic. I left before 7am today and it took me 90 mins.
Combination of bad weather and N7 roadworks make it a nightmare.
Fuck that is my future right there! Where are you going from - to? I expect to be on the N7 at Johnstown at 7am going to Tallaght.
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
Fuck that is my future right there! Where are you going from - to? I expect to be on the N7 at Johnstown at 7am going to Tallaght.
You'll be completely fine, you're on the right side of it. I live close to Kildare Village and I work in D15, it's approx 40 mins without traffic but around 90 mins when it's bad.
I drive home via Maynooth some days if I check the ETA and there is little difference between the two because I figure lower chance of traffic getting worse via Maynooth than M50/N7
A Population of Fish
There is a population of fish numbering 10000 at the start of the year.
Every month 5% of the fish alive at the start of the month, excluding offspring, die during the month.
For the months of May/June/July 25% of the fish alive at the start of the month pair off and pro-create. Each pair have 3 off-spring.
Every month, 15% of newly born fish over 3 months old depart the colony.
Assumptions:
Off spring do not pro-create
Make any other assumptions you need if you are in doubt.
Write code to calculate the number of fish at the end of the year
Every single person in here had a different answer
I figured I’d give this a go through excel but I can’t write code, so I’m just looking for the right answer
Chris Hughton should take it. Would easily double his current salary for far less work and will take him nicely to retirement. Realistically cant take Brighton much higher so would be a smart time to leave.
Fuck that is my future right there! Where are you going from - to? I expect to be on the N7 at Johnstown at 7am going to Tallaght.
Ooof. That's a large chunk of your day wasted sitting in traffic. Feeling better about location choices myself with my 11 min rail journey/20 min cycle.
I thought there was reasonably good commuter rail links into/out of Kildare.
"We are not Europeans. Those people on the continent are freaks."
My commute is getting steadily worse. I'm leaving earlier than ever but I can't seem to beat the traffic. I left before 7am today and it took me 90 mins.
Combination of bad weather and N7 roadworks make it a nightmare.
It will only be worse on the way into town when the N7 road works finish.
At the moment the works are operating as a filter slowing down traffic as it gets congested the whole way from Kill to the M50. Once it is finished it will only mean more gobshite lane weavers causing crashes as they run into traffic at higher speed.
On the positive side going south will improve though I expect there will still be that irrational accordion effect when its busy caused by cars slowing down dramatically when they stop seeing the road 500M ahead of them as they meet the small rise over the canal about 1500M from J9 (the ball) heading south.
Very frustrating that I'm rarely able to use the train despite it being very convenient for me
My commute is getting steadily worse. I'm leaving earlier than ever but I can't seem to beat the traffic. I left before 7am today and it took me 90 mins.
Combination of bad weather and N7 roadworks make it a nightmare.
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
Ooof. That's a large chunk of your day wasted sitting in traffic. Feeling better about location choices myself with my 11 min rail journey/20 min cycle.
I thought there was reasonably good commuter rail links into/out of Kildare.
Kildare Town station into Heuston is 28 mins on some trains in the morning so it is very good at times. I suspect they'll push the leap card out there or maybe as far as Portlaoise as it's currently to Naas/Sallins.
My 40 minute walk to work along the beach just got a little bit colder too. Was raining this morning too. Shit buzz
I always dismissed the value of living close to work/college until I was living in Ranelagh last year and then when I moved home i really noticed the difference. It's still hard to justify the rent though
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
I wrote some bad code in 5 minutes to do this, may be wrong but I'm usually right
SPOILER
Code:
# Can't have partial fish. Ignore floats, make it an int.
alive = 10000
d={}
for i in range (0,12):
if i == 4 or i == 5 or i == 6:
# approximate 12.5%
procreate = alive/8
offspring = procreate*3
d["offspring{0}".format(i)]=offspring
if i > 7:
d['offspring4'] = d['offspring4'] - d['offspring4']/100*15
if i > 8:
d['offspring5'] = d['offspring5'] - d['offspring5']/100*15
if i > 9:
d['offspring6'] = d['offspring6'] - d['offspring6']/100*15
dead = alive/20
alive = alive - dead
for i in d:
alive += d[i]
print alive
Rufio
Start of year,10000
Monthly rate of decline,0.05
May/June/July,0.125,Pair off and pro create, each pair have three offspring
Every Month 15% of newly born fish >3months depart
January,February,March,April,May,June,July,August,September,October,November,December
End of month total,9500,902...
I've put mine in, however I did not round figures to the nearest point....
Rufio
Start of year,10000
Monthly rate of decline,0.05
May/June/July,0.125,Pair off and pro create, each pair have three offspring
Every Month 15% of newly born fish >3months depart
January,February,March,April,May,June,July,August,September,October,November,December
End of month total,9500,902...
I've put mine in, however I did not round figures to the nearest point....
The difference between yours and mine is of semantics.
1) You do not round fish, I do.
2) You count fish born in a month as being one month old by the start of the next month, whereas I count them as not fully one month old, so our months are offset by one when it comes to counting age > 3 months.
We get the same answers if those are accounted for either way.
SPOILER
Code:
alive = float(10000)
d={}
for i in range (0,12):
if i == 4 or i == 5 or i == 6:
# approximate 12.5%
procreate = alive/8
offspring = procreate*3
d["offspring{0}".format(i)]=offspring
if i > 6:
d['offspring4'] = d['offspring4'] - d['offspring4']/100*15
if i > 7:
d['offspring5'] = d['offspring5'] - d['offspring5']/100*15
if i > 8:
d['offspring6'] = d['offspring6'] - d['offspring6']/100*15
dead = alive/20
alive = alive - dead
for i in d:
alive += d[i]
print alive
python fish.py
9966.44166297
I wrote some bad code in 5 minutes to do this, may be wrong but I'm usually right
SPOILER
Code:
# Can't have partial fish. Ignore floats, make it an int.
alive = 10000
d={}
for i in range (0,12):
if i == 4 or i == 5 or i == 6:
# approximate 12.5%
procreate = alive/8
offspring = procreate*3
d["offspring{0}".format(i)]=offspring
if i > 7:
d['offspring4'] = d['offspring4'] - d['offspring4']/100*15
if i > 8:
d['offspring5'] = d['offspring5'] - d['offspring5']/100*15
if i > 9:
d['offspring6'] = d['offspring6'] - d['offspring6']/100*15
dead = alive/20
alive = alive - dead
for i in d:
alive += d[i]
print alive
10820
oddly...
Code:
# Can't have partial fish. Ignore floats, make it an int.
alive = 10000
d={}
for i in range (0,12):
if i == 4 or i == 5 or i == 6:
# approximate 12.5%
procreate = alive/8
offspring = procreate*3
d["offspring{0}".format(i)]=offspring
if i > 7:
d['offspring4'] = d['offspring4'] - d['offspring4']/100*15
if i > 8:
d['offspring5'] = d['offspring5'] - d['offspring5']/100*15
if i > 9:
d['offspring6'] = d['offspring6'] - d['offspring6']/100*15
dead = alive/20
alive = alive - dead
for i in d:
alive += d[i]
print (alive)
in python gives me 10771.648860561916
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
USE tempdb
DECLARE @fish FLOAT = 10000;
DECLARE @offSpring FLOAT = 0;
DECLARE @departed FLOAT = 0;
DECLARE @month INT = '1';
WITH fish (
fish
,offSpring
,leavers
,[month]
)
AS (
SELECT @fish AS fish
,@offSpring AS offSpring
,@departed AS leavers
,@month AS [month]
UNION ALL
SELECT ROUND(fish.fish - (fish.fish * 0.05), 0)
,fish.offSpring + CASE
WHEN fish.[month] IN (
1
,2
,3
,7
,8
,9
,10
,11
,12
)
THEN 0
ELSE ROUND(ROUND(ROUND(fish.fish - (fish.fish * 0.05), 0) * 0.125, 0) * 3, 0)
END
,fish.leavers + CASE
WHEN fish.[month] IN (
1
,2
,3
,4
,5
,6
)
THEN 0
ELSE ROUND(ROUND(ROUND(ROUND(fish.fish - (fish.fish * 0.05), 0) * 0.125, 0) * 3, 0) * 0.15, 0)
END AS leavers
,fish.[month] + 1 AS [month]
FROM fish
WHERE [month] < 13
)
--Select * from fish -- Array of fish*/
-- max val
SELECT (fish.fish + fish.offSpring) - fish.leavers AS colonysize
FROM fish
WHERE fish.[month] = 13;
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
Rufio
Start of year,10000
Monthly rate of decline,0.05
May/June/July,0.125,Pair off and pro create, each pair have three offspring
Every Month 15% of newly born fish >3months depart
January,February,March,April,May,June,July,August,September,October,November,December
End of month total,9500,902...
I've put mine in, however I did not round figures to the nearest point....
# Can't have partial fish. Ignore floats, make it an int.
alive = 10000
d={}
for i in range (0,12):
if i == 4 or i == 5 or i == 6:
# approximate 12.5%
procreate = alive/8
offspring = procreate*3
d["offspring{0}".format(i)]=offspring
if i > 7:
d['offspring4'] = d['offspring4'] - d['offspring4']/100*15
if i > 8:
d['offspring5'] = d['offspring5'] - d['offspring5']/100*15
if i > 9:
d['offspring6'] = d['offspring6'] - d['offspring6']/100*15
dead = alive/20
alive = alive - dead
for i in d:
alive += d[i]
print (alive)
in python gives me 10771.648860561916
You'll be using python 3 so, I'm on python 2 where division works differently (/ operator is integer division if inputs are integers.). You need to explicitly cast to an int at the start to get my answer.
Of course it being python I can get your answer by looking into the future by just adding this line:
I assumed 25% coupling off meant 25% male and 25% female, without that its 12,089
Make any other assumptions you need if you are in doubt.
I assume some fish adopted and there was a fish serial killer to get my answer
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
You'll be using python 3 so, I'm on python 2 where division works differently (it rounds). You need to explicitly cast to an int at the start to get my answer.
I'm very much a python newb. Just thought it was interesting
had to put parentheses around alive too to run it.
People say I should be more humble I hope they understand, they don't listen when you mumble
Get a shiny metal Revolut card! And a free tenner! https://revolut.com/referral/jamesb8!G10D21
I'm very much a python newb. Just thought it was interesting
had to put parentheses around alive too to run it.
Yeah, it changed into a function only in python 3, which allows some more magic to be accomplished than before. I must think about making better code than mine as an exercise.
Chris Hughton should take it. Would easily double his current salary for far less work and will take him nicely to retirement. Realistically cant take Brighton much higher so would be a smart time to leave.
Mostly because I used to live nearby-ish and remember it being there. Every time I pass the derelict site now I try to remember the name of the pub but couldn't really. That's all there is to it.
If I present a representation of a fake news event to you and you say you remember seeing or hearing that then it's a false memory. It's actually an interesting area, especially in respect of false autobiographical memories where people are induced to falsely remember, say, a childhood hot air balloon ride.
I can't remember how to post an image here, but if I could, it would be an image of Hannibal Lecter.
Official Head Marshall of Waterford Gay Pride Festival 2015
We process personal data about users of our site, through the use of cookies and other technologies, to deliver our services, personalize advertising, and to analyze site activity. We may share certain information about our users with our advertising and analytics partners. For additional details, refer to our Privacy Policy.
By clicking "I AGREE" below, you agree to our Privacy Policy and our personal data processing and cookie practices as described therein. You also acknowledge that this forum may be hosted outside your country and you consent to the collection, storage, and processing of your data in the country where this forum is hosted.
Comment