Merge branch 'master' of https://github.com/tqchen/xgboost

README.md

@@ -4,7 +4,7 @@ An efficient general purpose gradient boosting (tree) library.
 
 Creater: Tianqi Chen
 
-Documentation: https://github.com/tqchen/xgboost/wiki
+Turorial and Documentation: https://github.com/tqchen/xgboost/wiki
  
 
 Features
@@ -15,7 +15,7 @@ Features
   - Efficient implementation that optimizes memory and computation.
 * Layout of gradient boosting algorithm to support generic tasks, see project wiki.
 
-Planned key components
+Supported key components
 =======
 * Gradient boosting models: 
     - regression tree (GBRT)
@@ -23,10 +23,14 @@ Planned key components
 * Objectives to support tasks: 
     - regression
     - classification
+* OpenMP implementation
+
+Planned components
+=======
+* More objective to support tasks: 
     - ranking
     - matrix factorization
     - structured prediction
-(3) OpenMP implementation
 
 File extension convention: 
 (1) .h are interface, utils and data structures, with detailed comment; 

demo/test/runexp.sh

@@ -4,23 +4,20 @@ python mapfeat.py
 # split train and test
 python mknfold.py agaricus.txt 1
 # training
-../../xgboost mushroom.conf num_round=1 model_out=full.model bst:max_depth=3
+../../xgboost mushroom.conf num_round=2 model_out=full.model bst:max_depth=3
 ../../xgboost mushroom.conf task=dump model_in=full.model fmap=featmap.txt name_dump=dump.full.txt
 
-# constrain
+# major element of batch running: add batch prefix to each setting, batch:run=1 will run that action
-../../xgboost mushroom.conf num_round=1 model_out=ban.model bst:max_depth=3 bst:fban=22-31 
 
-# constrain
+../../xgboost mushroom.conf model_in=full.model model_out=m1.model task=interact\
-../../xgboost mushroom.conf num_round=1 model_out=pass.model bst:max_depth=3 bst:fdefault=-1 bst:fpass=22-31
+ batch:interact:booster_index=0 batch:bst:interact:remove=1 batch:run=1\
+ batch:interact:booster_index=1 batch:bst:interact:remove=1 batch:run=1\
+ batch:interact:booster_index=1 batch:bst:interact:expand=9 batch:run=1\
 
-../../xgboost mushroom.conf task=dump model_in=ban.model fmap=featmap.txt name_dump=dump.ban.txt
+../../xgboost mushroom.conf task=dump model_in=m1.model fmap=featmap.txt name_dump=dump.m1.txt
-../../xgboost mushroom.conf task=dump model_in=pass.model fmap=featmap.txt name_dump=dump.pass.txt
 
 echo "========full======="
 cat dump.full.txt
 
-echo "========ban======="
+echo "========m1======="
-cat dump.ban.txt
+cat dump.m1.txt
-
-echo "========pass======="
-cat dump.pass.txt

regression/xgboost_reg_main.cpp

@@ -73,6 +73,9 @@ namespace xgboost{
                 if( !strcmp("name_pred",  name ) )   name_pred = val;
                 if( !strcmp("dump_stats", name ) )   dump_model_stats = atoi( val );
                 if( !strcmp("interact:action",  name ) )  interact_action = val;
+                if( !strncmp("batch:",  name, 6 ) ){
+                    cfg_batch.PushBack( name + 6, val );
+                }
                 if( !strncmp("eval[",  name, 5 ) ) {
                     char evname[ 256 ];
                     utils::Assert( sscanf( name, "eval[%[^]]", evname ) == 1, "must specify evaluation name for display");
@@ -167,13 +170,27 @@ namespace xgboost{
             inline void TaskInteractive( void ){
                 const time_t start    = time( NULL );
                 unsigned long elapsed = 0;
+                int batch_action = 0;
+                
+                cfg_batch.BeforeFirst();
+                while( cfg_batch.Next() ){
+                    if( !strcmp( cfg_batch.name(), "run" ) ){
                         learner.UpdateInteract( interact_action );
+                        batch_action += 1;
+                    } else{
+                        learner.SetParam( cfg_batch.name(), cfg_batch.val() );
+                    }
+                }
+
+                if( batch_action == 0 ){
+                    learner.UpdateInteract( interact_action );
+                }
                 utils::Assert( model_out != "NULL", "interactive mode must specify model_out" );
                 this->SaveModel( model_out.c_str() );
                 elapsed = (unsigned long)(time(NULL) - start); 
 
                 if( !silent ){
-                    printf("\ninteractive update, %lu sec in all\n", elapsed );
+                    printf("\ninteractive update, %d batch actions, %lu sec in all\n", batch_action, elapsed );
                 }
             }
 
@@ -245,6 +262,8 @@ namespace xgboost{
             std::vector<std::string> eval_data_names;            
             /*! \brief saves configurations */
             utils::ConfigSaver cfg;
+            /*! \brief batch configurations */
+            utils::ConfigSaver cfg_batch;
         private:
             DMatrix data;
             std::vector<DMatrix*> deval;