Data-mining and Machine Learning approach to trading?

Discussion in 'Data Sets and Feeds' started by mizhael, Jun 28, 2009.

  1. Hi all,

    I am building my programming trading system using InteractiveBrokers API.

    I am thinking of the data-mining and machine learning approach. Could anybody give me some pointers about successful strategies that utilize data-mining and machine learning techniques?

    Any online resources, online communities that dedicate to the data-mining and machine learning approach to trading?

    Thanks a lot!
     
  2. I have noticed that just about every single questionnaire thread you have posed is a one liner query on some hot topic of the day with numerous different topics, yet, you never add anything to the discussion.

    Honestly, what kind of incentive are you given to generate activity here? Monetary? How Much? Where do you find your keyword list of topics to question?

    It sounds as if you already have some experience in data mining! In fact, you almost seem to pass the Turing test with flying colors. Perhaps you should be doing the informing.


    Thanks a lot!
     
  3. Previously I have posted an academic paper and given my comment about the paper. Have you seen that one?

    I say what I know and ask what I don't know, what's wrong with that?

    Why do you need to question one's incentive? I can ascertain you that I don't receive money for asking questions. And I don't think ET need to give money to people to generate traffic.

    By questioning other people's incentive about posting questions, do you think you add positive value to the forum? Why don't you focus on the knowledge and trading, instead of raising personal attack and questioning other people's incentive?

    Finally, of course I am interested and asking about the hot topic... what's wrong with that?
     
  4. No, you don't get it. I am doing precisely what you advocate. I am gathering knowledge by asking questions. What is wrong with that? Honestly, I'm ascertained! I think I even caught that academic paper.

    I truly wanted to understand if you really actually care about your questions, or have some other reason for posing them. Your history suggest the latter, therefore I was curious to know why. In fact, quite honestly, I am impressed by your ability to filter sparse interesting keyword related topics on such a wide range of areas.

    Is there anything wrong with that? Also, that was a surprisingly fast reply. Almost... automated?:D What do you think about the Turing test as applied to trading forums?
     
  5. rosy2

    rosy2

    do you even know what "data-mining" and "machine learning" mean? I doubt it.

    stanford has posted a machine learning class on youtube.

    http://www.youtube.com/watch?v=UzxYlbK2c7E
     

  6. Yes, I have watched the video...
     
  7. Hi Mizhael,

    Now that you have watched the video, what parts do you find relevant to trading? Did they cover Searle Chinese wall or Turing test? Which do you think applies more to a trading forum?

    Thanks!
     
  8. Murray Ruggiero

    Murray Ruggiero Sponsor

    Exactly what are you trying to do. Do you want to create trading rules from data. Train a model to predict price change or identify patterns ?.

    Genetic algorithms or a technology call Rough Sets can be used create trading rules from data. Genetic programming also can be used to mine rules from data.

    Neural Networks or Kernal regression can be used to create prediction models, Kernal regression can also be used for pattern detection.
     
  9. Data-mining and IB? Oxymoron?
     
  10. I think you are missing the point of my replies. I don't think he's looking for data mining for trading; rather, I think he's mining for clicks and response activity (i.e. board shill).:D
     
    #10     Jul 2, 2009