Machine learning algorithm in Minorthird

  • Prakash

    Prakash - 2011-11-02

    Dear All,

    I need your help

    I am writing all 10 algorithms of machine learning that are in minorthird. Can anybody say me the full-form of each of them. I listed down according to the javadoc of minothird, but I may be wrong. So please can you help me to get the right and also it would be great if you tell me the reference from where you know these all.

    VPSMMLearner2                        - Voted Perception Hidden Semi Markov Model Learner 2

    VPCMMLearner                          - Voted Perception Conditional Markov Model

    SVCMMLearner                          - Support vector machine model learner

    MEMMLearner                            - Maxent Markov Model

    CRFAnnotatorLearner               - Conditional Random Field

    SegmentAnnotatorLearner       - Segment Annotator Learner

    SequenceAnnotatorLearner    - Sequence Annotator Learner

    SemiCRFAnnotatorLearner     - Semi Conditional Random Field

    VPHMMLearner                          - Voted Perception Hidden Markov Model

    VPSMMLearner                          - Voted Perception Hidden Semi Markov Model

  • Frank Lin

    Frank Lin - 2011-11-02

    I think these are correct:

    VPSMM - voted perceptron semi-Markov model
    VPCMM - voted perceptron conditional Markov model
    SVCMM - support vector conditional Markov model
    MEMM - maximum entropy Markov model
    CRF - conditional random fields
    SemiCRF - semi-conditional random fields
    VPHMM - voted perceptron hidden Markov model
    VPSMM - voted perceptron semi-Markov model

  • Prakash

    Prakash - 2011-11-02

    Thank you so much Frank,

    but what about

    SegmentAnnotatorLearner - Segment Annotator Learner
    SequenceAnnotatorLearner - Sequence Annotator Learner

    what type of algorithm are they

    and also you wrote  VPSMM - voted perceptron semi-Markov model two times. Is it  because of VPSMMLearner2 and VPSMMLearner and what are the difference between them.

  • Frank Lin

    Frank Lin - 2011-11-02

    Sequence and segment learners are not algorithms; they are a general class of models that do not assume each example is independent of others; instead they take into account the order in which they appear in a dataset.

    Most annotator learners in minorthird are sequence or segment annotators.

    For more details on VPSMMLearner2 please refer to the javadoc at:

    Look up Recommended.VPSMMLearner2

  • Prakash

    Prakash - 2011-11-12


    Thanks for your continues reply, I do have so many dought I will keep on writing to you in this open forum.

    Right now I do have question as why these algorithm are called voted perception. Do we need to right all the "voted perceptron semi-Markov model" what actually is voted perception means??

  • Frank Lin

    Frank Lin - 2011-11-13

    Look up Recommended.VPSMMLearner2 in the javadoc - there is a paper reference, which will give you more detailed information on voted perceptron with semi-markov model.


Log in to post a comment.