Changes between Version 4 and Version 5 of MultiLabelClassification


Ignore:
Timestamp:
05/16/11 18:40:27 (3 years ago)
Author:
wencanluo
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • MultiLabelClassification

    v4 v5  
    1414 
    1515=== Method 3: adding a special value into their 'attributes' dictionary ===  
     16 
     17The Variable descriptors can store additional variables, see [//doc//orange25/orange.data.feature.html#attributes Storing additional variables]. So we can add a special value, like "lable" into the 'attributes' dictionary with vaule 1. In addition, we can set the additions to meta type, so that such labels would not be processed by the old one-label methods. 
     18 
     19In this way, a tab file will look like below: 
     20{{{ 
     21Feature Sports  Religion        SCience Politics 
     22d       d       d       d       d 
     23        m label=1       m label=1       m label=1       m label=1 
     241       1       0       0       1 
     252       0       0       1       1 
     263       1       0       0       0 
     274       0       1       1       0 
     28}}} 
     29 
     30where 'Sports','Religion','SCience','Politics' are multi labels.The first example has feature "Feature=1" and belongs to label 'Sports' and label 'Politics'; The first example belongs to label 'SCience' and label 'Politics'; The third example belongs to only label 'Sports'; The fourth example belongs to label 'Religion' and label 'SCience'; 
     31 
     32What needs to be changed? 
     33* Getclass method 
     34 
     35What doesn't need to be changed? 
     36* Input file parser, like Tab file 
     37* one-label methods 
    1638 
    1739=== Method 4: allow to multi 'class' optional flags ===  
     
    5173See [//doc/widgets Widget development] manual 
    5274 
    53 == Timeline == 
    54  April 25 – May 23 (Before official coding time) :: To discuss the details about my ideas with my mentor to archive a final agreement, including designing the dataset support, choosing which transformation and adaptive methods to implement. Familiar with structure of Orange source code. How the Python code and C++ code combined? Based on the final agreement, I will write some testing code to make clear about all goals. 
    55  May 23 – June 18 (Official coding period starts) :: Start to design the framework to support multi-label classification, including the multi-label data structure –instance, instances, attribute, evaluator, etc. 
    56  Coding on designing basic multi-label dataset, two of problem-transformation methods-Binary relevance (BR), Calibrated label ranking (CLR), one GUI widget, and two evaluation measures: Example-based Hamming-Loss, Classfication Accuracy, Precision, Recall; Label-based. 
    57  June 18 – July 5 :: Finish the work on improving problem-transformation methods, and test the whole work to ensure it can work properly. 
    58  Start to code on algorithm adaptation method: ML-KNN, Multi-class multi-label perceptron (MMP). 
    59  July 6 – July 15 (Mid-term) :: Finish the work on adaptation models and do some test work to ensure it could work properly. Start to implement feature selection methods: LP based and Transformation based. 
    60  Submit mid-term evaluation. 
    61  July 16– July 31 :: Finish all my work on the Multi-Label project and do bug fixing work and test. 
    62  Make a document about what we have now and what to do next. 
    63  August 1 – August 15 :: Redundant time for some unpredictable stuff to do. 
    64  If it is possible, I could work on to implement more problem-transformation models, adapted models and evaluation methods. 
    65  Submit final evaluation. 
     75== ToDo List == 
     76Before May 23 (official coding time) 
     77* Designing the dataset support 
     78* choosing which transformation and adaptive methods to implement 
     79* Familiar with structure of Orange source code 
     80* How the Python code and C++ code combined?  
    6681 
     82May 23 – June 18 (Official coding period starts) 
     83* design the framework to support multi-label classification, including the multi-label data structure –instance, instances, attribute, evaluator, etc. 
     84* Coding on designing basic multi-label dataset 
     85* two of problem-transformation methods-Binary relevance (BR), Calibrated label ranking (CLR) 
     86* one GUI widget 
     87* two evaluation measures: Example-based Hamming-Loss, Classfication Accuracy, Precision, Recall; Label-based. 
     88* convert [http://mulan.sourceforge.net/format.html mulan multilabel data file format] to Tab file 
     89 
     90June 18 – July 5  
     91* Finish the work on improving problem-transformation methods 
     92* test the whole work to ensure it can work properly. 
     93* Start to code on algorithm adaptation method: ML-KNN, Multi-class multi-label perceptron (MMP). 
     94 
     95July 6 – July 15 (Mid-term)  
     96* Finish the work on adaptation models and do some test work to ensure it could work properly 
     97* Start to implement feature selection methods: LP based and Transformation based. 
     98* Submit mid-term evaluation. 
     99July 16– July 31 
     100* Finish all my work on the Multi-Label project and do bug fixing work and test. 
     101* Make a document about what we have now and what to do next. 
     102August 1 – August 15  
     103* Redundant time for some unpredictable stuff to do. 
     104* Submit final evaluation. 
     105