The Hero with a Thousand Faces -
Why do we keep telling the same stories over and over?
|
Saturday, 21 January 2017
quote [ Thanks to the internet, the American language seems to be exploding at an almost unclockable pace. Polyglot social media has exposed American English, with its historically promiscuous embrace of new idioms, to more digital pidgins, foreign words, microdialects, pictograms, neologisms, and cryptic symbols than any one language user can ever expect to brook. ]
This is mostly a test of the "prep post" system. I hope you find it interesting, anyway.
|
lilmookieesquire said @ 8:48pm GMT on 21st January
Nothing too wonderful. I had issues with data size and cleaning data that made meaningful results difficult to get to (the Stanford paper which is better was a thesis so they had a good chunk of time, I had like two or three weeks and very limited space) but being positive was generally slightly more beneficial.
The next thing I would have investigated was the time of the request.
The real difficulty was that most of the accounts were a first comment in that sub account.
My paper was for a data mining and analysis class so the paper was more of a demonstration of techniques and cleaning the data took so. Much. Time.
In my defense my professor did one on mortgages and it took her about a year to clean the data and make it useable. The nice thing is that even those tools are advancing at great speed.
SPSS (for example) is so so much easier to use now than it was in 2008. Everything is getting more automated but the tricky part is understanding the techniques and assumptions in that automation (which is something that lead to the financial meltdown with AAA rated bonds containing huge amounts of risk via inability of clients to pay back the loans.
So there's real danger in just being a button pusher without understanding the mechanics- but we're all guilty of that to some degree.
lilmookieesquire said @ 8:50pm GMT on 21st January
Nothing too wonderful. I had issues with data size and cleaning data that made meaningful results within my time and scope limits difficult to get to (the Stanford paper which is better was a graduate thesis so they had a good chunk of time, vs I had like two or three weeks and very limited space which included picking and understanding the data set) but being positive was generally slightly more beneficial vs the words "not" or "don't" being more associated with failed requests. (The word truncation function I was using was also imperfect)
The next thing I would have investigated was the time of the request.
The real difficulty was that most of the accounts were a first comment in that sub account.
My paper was for a data mining and analysis class so the paper was more of a demonstration of techniques and cleaning the data took so. Much. Time.
In my defense my professor did one on mortgages and it took her about a year to clean the data and make it useable. The nice thing is that even those tools are advancing at great speed.
SPSS (for example) is so so much easier to use now than it was in 2008. Everything is getting more automated but the tricky part is understanding the techniques and assumptions in that automation (which is something that lead to the financial meltdown with AAA rated bonds containing huge amounts of risk via inability of clients to pay back the loans.
So there's real danger in just being a button pusher without understanding the mechanics- but we're all guilty of that to some degree.
lilmookieesquire said @ 8:52pm GMT on 21st January
Nothing too wonderful. I had issues with data size and cleaning data that made meaningful results within my time and scope limits difficult to get to (the Stanford paper which is better was a graduate thesis so they had a good chunk of time, vs I had like two or three weeks and very limited space which included picking and understanding the data set) but being positive was generally slightly more beneficial vs the words "not" or "don't" being more associated with failed requests. (The word truncation function I was using was also imperfect)
Obviously upvotes correlates with successful requests but that would have been an entirely different paper - but was a "with more funding..."; "future research" recommendation.
The next thing I would have investigated was the time of the request.
The real difficulty was that most of the accounts were a first comment in that sub account.
My paper was for a data mining and analysis class so the paper was more of a demonstration of techniques and cleaning the data took so. Much. Time.
In my defense my professor did one on mortgages and it took her about a year to clean the data and make it useable. The nice thing is that even those tools are advancing at great speed.
SPSS (for example) is so so much easier to use now than it was in 2008. Everything is getting more automated but the tricky part is understanding the techniques and assumptions in that automation (which is something that lead to the financial meltdown with AAA rated bonds containing huge amounts of risk via inability of clients to pay back the loans.
So there's real danger in just being a button pusher without understanding the mechanics- but we're all guilty of that to some degree.
<-- Entry / Current Comment
lilmookieesquire said @ 8:48pm GMT on 21st January [Score:1 Interesting]
Nothing too wonderful. I had issues with data size and cleaning data that made meaningful results within my time and scope limits difficult to get to (the Stanford paper which is better was a graduate thesis so they had a good chunk of time, vs I had like two or three weeks and very limited space which included picking and understanding the data set) but being positive was generally slightly more beneficial vs the words "not" or "don't" being more associated with failed requests. (The word truncation function I was using was also imperfect)
Obviously upvotes correlates with successful requests but that would have been an entirely different paper - but was a "with more funding..."; "future research" recommendation.
The next thing I would have investigated was the time of the request.
The real difficulty was that most of the accounts were a first comment in that sub account.
My paper was for a data mining and analysis class so the paper was more of a demonstration of techniques and cleaning the data took so. Much. Time.
In my defense my professor did one on mortgages and it took her about a year to clean the data and make it useable. The nice thing is that even those tools are advancing at great speed.
SPSS (for example) is so so much easier to use now than it was in 2008. Everything is getting more automated but the tricky part is understanding the techniques and assumptions in that automation (which is something that lead to the financial meltdown with AAA rated bonds containing huge amounts of risk via inability of clients to pay back the loans.
So there's real danger in just being a button pusher without understanding the mechanics- but we're all guilty of that to some degree.
