home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   comp.ai      Awaiting the gospel from Sarah Connor      1,954 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 820 of 1,954   
   Predictor to albinali   
   Re: Hidden Markov Models   
   27 Oct 05 15:33:17   
   
   From: predictr@bellatlantic.net   
      
   albinali wrote:   
   > I need some feedback/suggestions on how to approach the following   
   > problem using Hidden Markov Models:   
   >   
   > For a particular application on a desktop (e.g. an FTP client), if we   
   > monitor the sequence of mouse events (particularly mouse clicks and   
   > their coordinates), can we determine if the application will attempt to   
   > transfer data on a wireless card.  For example, one would think that   
   > using specific menus on a particular FTP client would initiate an FTP   
   > send/get request and therefore we may be able to predict such events   
   > before the data is relayed to the wireless card.   
   >   
   > Thus, my goal is to predict the state of an application whether it is   
   > sending data or not.   
   >   
   > To tackle this problem, I am planning to use HMMs. I plan to have at   
   > least a 2 -state HMM (with one state signfying that the application is   
   > not in a sending mode and another state indicating that the application   
   > is in a sending mode). The observations that I will gather are mouse   
   > clicks associated with the activity of the wireless card (i.e. if there   
   > was data sent during the mouse clicks or not).   
      
      
   I'm not sure why you selected HMM, specifically, but I would think that   
   you'd find that either: 1. there are so few states that a simple   
   probability lookup table could be constructed, or 2. there are too many   
   states for a simple table or you wish to include other information,   
   indicating the use of some sort of statistical (naive Bayes, CART,   
   etc.) or machine learning (neural network, rule induction) solution.   
      
      
   -Will Dwinnell   
   http://will.dwinnell.com   
      
   [ comp.ai is moderated.  To submit, just post and be patient, or if ]   
   [ that fails mail your article to , and ]   
   [ ask your news administrator to fix the problems with your system. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca