Pages:
Author

Topic: Wall Observer Observer - MtGoxUSD wall movement tracker tracker (Read 4836 times)

legendary
Activity: 1904
Merit: 1002
Ok, so I have important stuff to do but keep procrastinating reading bitcoin stuff so I may as well improve this. I need a list of bullish and bearish terms to create a sentiment index for that thread.

Classify days as bullish/bearsish/neutral.  Use TF-IDF to pick out the most frequent terms on bullish or bearish days.

Do you mean just choose a few as training samples and subjectively classify them, or is there a specific clustering algorithm that is good for this type of thing? It's all new to me.

1. Classify time period as bullish/bearish/neutral based on price action

2.  Gather statistics on word frequency per time period as well as word frequencies from a general text corpus.

3.  Use statistics to compute the frequency of each word for each class (bullish, bearish, neutral).  These are your term frequencies for that class.  Weight each term frequency by the inverse of the document frequency (a bit of terminology here: document frequency in this context refers to the frequency of the word in the general text corpus).  This weighting helps filter out common words and accentuate words that are unique to the target class.

4.  Pick the top few TF-IDF words to predict for each class.

hero member
Activity: 728
Merit: 500
Ok, so I have important stuff to do but keep procrastinating reading bitcoin stuff so I may as well improve this. I need a list of bullish and bearish terms to create a sentiment index for that thread.

Classify days as bullish/bearsish/neutral.  Use TF-IDF to pick out the most frequent terms on bullish or bearish days.

Do you mean just choose a few as training samples and subjectively classify them, or is there a specific clustering algorithm that is good for this type of thing? It's all new to me.
hero member
Activity: 728
Merit: 500
I gave up on the sentiment index for now, but here is a new type of plot. The diameter of each "Bead on a VWAP" corresponds to the number of posts by either Proudhon (who has been bearish this whole time if I remember correctly), and Loaded, who seems to post before interesting things occur. Also the volume and posts per day charts are now of the same type.

hero member
Activity: 728
Merit: 500
Yes, also, anyone want to help on the regex to extract only the non-quote text from this:
Code:
a corresponding timesstamp would be nice

Hmmm... aren't the messages on the forum time-stamped already?

Yes, but he posted images that he'd taken hours or days before he started the thread, I believe.

yes the first 2 do not have a correct times stamp but are relatively new

the time from the posts should be fine
I wonder why turning on the "show volume" (the option that darkens the green line anywhere the volume spikes) isn't simply on by default.  It's extra information at a cost of zero screen clutter.

thanks i was forgetting about this option

Its this post:
a corresponding timesstamp would be nice

Hmmm... aren't the messages on the forum time-stamped already?

Yes, but he posted images that he'd taken hours or days before he started the thread, I believe.

yes the first 2 do not have a correct times stamp but are relatively new

the time from the posts should be fine
I wonder why turning on the "show volume" (the option that darkens the green line anywhere the volume spikes) isn't simply on by default.  It's extra information at a cost of zero screen clutter.

thanks i was forgetting about this option
legendary
Activity: 1904
Merit: 1002
I need a list of bullish and bearish terms to create a sentiment index for that thread.

Could you give us a list of the top-100 most frequently used words in the thread? (to start us off. we can then eliminate unrelated terms.)

By the way, can you go a little further back in the graph? (if it's not too much trouble. To compare 2013 to 2011.)

Edit: Plus what notme posted

Classify days as bullish/bearsish/neutral.  Use TF-IDF to pick out the most frequent terms on bullish or bearish days.

Yes, I can do the word frequency list, this is actually gunna be a bit harder then I first thought due to the quoting. I need to figure out how to filter those. I cannot go back any further because the thread did not exist before June 2012.

Good point.  You might be short on bearish examples.
hero member
Activity: 728
Merit: 500
I need a list of bullish and bearish terms to create a sentiment index for that thread.

Could you give us a list of the top-100 most frequently used words in the thread? (to start us off. we can then eliminate unrelated terms.)

By the way, can you go a little further back in the graph? (if it's not too much trouble. To compare 2013 to 2011.)

Edit: Plus what notme posted

Classify days as bullish/bearsish/neutral.  Use TF-IDF to pick out the most frequent terms on bullish or bearish days.

Yes, I can do the word frequency list, this is actually gunna be a bit harder then I first thought due to the quoting. I need to figure out how to filter those. I cannot go back any further because the thread did not exist before June 2012.
full member
Activity: 210
Merit: 100
I need a list of bullish and bearish terms to create a sentiment index for that thread.

Could you give us a list of the top-100 most frequently used words in the thread? (to start us off. we can then eliminate unrelated terms.)

By the way, can you go a little further back in the graph? (if it's not too much trouble. To compare 2013 to 2011.)

Edit: Plus what notme posted

Classify days as bullish/bearsish/neutral.  Use TF-IDF to pick out the most frequent terms on bullish or bearish days.
legendary
Activity: 1904
Merit: 1002
Ok, so I have important stuff to do but keep procrastinating reading bitcoin stuff so I may as well improve this. I need a list of bullish and bearish terms to create a sentiment index for that thread.

Classify days as bullish/bearsish/neutral.  Use TF-IDF to pick out the most frequent terms on bullish or bearish days.
hero member
Activity: 728
Merit: 500
Ok, so I have important stuff to do but keep procrastinating reading bitcoin stuff so I may as well improve this. I need a list of bullish and bearish terms to create a sentiment index for that thread.
hero member
Activity: 728
Merit: 500



R code:
Code:
require(RCurl)

last.page<-15640
num<-"00"
allposts=NULL
while(as.numeric(num)<=last.page){
 
  forum<-getURL(
    paste("https://bitcointalk.org/index.php?topic=85687.",num, sep=""),
    ssl.verifypeer = FALSE, useragent = "R"
  )
 
 
  forum<-readLines(tc <- textConnection(forum)); close(tc)
  dates1<-grep('
',forum)
  dates2<-dates1[seq(2,length(dates1),by=2)]
  forum2<-forum[dates2]
  if(length(grep("edit",forum2))>0){
    forum3<-forum2[-grep("edit",forum2)]
  }else{
    forum3<-forum2
  }
 
  dates=NULL
  for(i in 1:length(forum3)){
    temp<-unlist(strsplit(forum3[i],">"))
    date<-strsplit(temp[2],"   
    dates<-rbind(dates,date)
  }
 
  allposts<-rbind(allposts,dates)
 
  num<-20 + as.numeric(num)
  print(num)
}

rownames(allposts)<-1:length(allposts)

allposts<-allposts[-which(allposts=="
alldates=NULL
for(i in 1:length(allposts)){
  temp<-unlist(
    strsplit(
      unlist(strsplit(allposts[i],", ")),
      " "
    )
  )
 
  temp2<-cbind(temp[1],temp[2],temp[3])
 
  alldates<-rbind(alldates,temp2)
}
alldates<-as.data.frame(alldates)
alldates[,1]<-as.character(alldates[,1])
alldates[,2]<-as.numeric(as.character(alldates[,2]))
alldates[,3]<-as.numeric(as.character(alldates[,3]))



mnth<-month.name[1]
yr<-2012
counts=NULL

for(yr in c(2012,2013)){
  for(m in 1:length(month.name)){
    mnth<-month.name[m]
    temp<-alldates[which(alldates[,1]==mnth & alldates[,3]==yr),]
    if(nrow(temp)>0){
     
     
      for(i in 1:max(temp[,2])){
        cnt<-length(which(temp[,2]==i))
       
        counts.temp<-cbind(unique(temp[,1]),i,unique(temp[,3]),cnt)
        counts<-rbind(counts,counts.temp)
      }
     
    }
   
  }
 
}

counts<-as.data.frame(counts)
counts[,1]<-as.character(counts[,1])
counts[,2]<-as.numeric(as.character(counts[,2]))
counts[,3]<-as.numeric(as.character(counts[,3]))
counts[,4]<-as.numeric(as.character(counts[,4]))



x.vals<-round(seq(1,nrow(counts), length=10),0)
x.labs<-paste(counts[x.vals,1],counts[x.vals,2], counts[x.vals,3])

plot(counts[,4], type="l", lwd=2, col="Red",
     ylab="Number of Posts",xaxt="n", xlab="",
     main="Posts per Day in Wall Observer Thread"
)
axis(side=1, at=x.vals, labels=x.labs)


hero member
Activity: 728
Merit: 500
Its running now... I didn't save the data the first time so it will take a bit.
sr. member
Activity: 252
Merit: 250
hero member
Activity: 728
Merit: 500
I've been thinking about "business models" that can fund research not backed by violence (patents and government funding). This thread is obviously a joke, but I want to try something out. If, after this date, a total of 0.1 bitcoins is sent to:

125x7TCMH7jxVPCiYAaaZUNbJiLpzEzMiR

I will post a new wall observer observer chart and post the code so anyone can do it.
sr. member
Activity: 252
Merit: 250
hero member
Activity: 509
Merit: 564
"In Us We Trust"
Thread fork!

Aww man, my last 3 posts in the other thread are basically worthless.

No they're just effectively worth half as much. They're still legitimate on both threads don't worry.
hero member
Activity: 544
Merit: 500
Thread fork!

Shit now you better do a graph of the graph of the growth of this thread vs. BTC price, vs. volatility. Then we can see which thread is the fastest generating fork.  Perhaps we'll need to downgrade to version 0.7.2 or earlier of this thread to keep the old forking wall tracker thread alive.

Maybe then in a few weeks we'll all have to upgrade to a new more stable wall tracker thread.   Wink
legendary
Activity: 1036
Merit: 1000
Wow, great job! Makes me wish I had hacker skillz.
legendary
Activity: 1638
Merit: 1001
₪``Campaign Manager´´₪
Seems to me both volatility and price are correlated with the number of posts.
(and I would expect sudden volatility after long period of quiet to have larger effect)
legendary
Activity: 1106
Merit: 1001
Thread fork!

Aww man, my last 3 posts in the other thread are basically worthless.

Uber confusing, but I Lol'ed when I realised what was happening.
legendary
Activity: 896
Merit: 1001
Thread fork!

Aww man, my last 3 posts in the other thread are basically worthless.
Pages:
Jump to: