merge develop

doug919 · Feb 13, 2015 · 3dc7944 · 3dc7944
2 parents 9033099 + 2abca90
commit 3dc7944
Show file tree

Hide file tree

Showing 2 changed files with 83 additions and 14 deletions.
diff --git a/README.md b/README.md
@@ -7,7 +7,75 @@ Python modules for analyzing LJ40K emotion data
 
 ![feelit flow](https://cloud.githubusercontent.com/assets/1659204/5698196/fd3873e8-9a42-11e4-803e-81c59a12c143.png)
 
+## batch/batchSimpleTrain.py
 
+perform SVM training for LJ40K
+
+1. usage
+
+	```
+	batchSimpleTraining.py [-h] [-k NFOLD] [-o OUTPUT_NAME] [-e EMOTION_IDS] 
+							[-c C] [-g GAMMA] [-s SCORE_DIR] [-v] [-d] 
+							feature_list_file
+
+	positional arguments:
+  		feature_list_file   This program will fuse the features listed in this
+                        	file and feed all of them to the classifier. The file
+                        	format is in JSON. See "feautre_list_ex.json" for
+                        	example
+
+	optional arguments:
+		-h, --help          show this help message and exit
+  		-k NFOLD, --kfold NFOLD
+                        	k for kfold cross-validtion. If the value less than 2,
+                        	we skip the cross-validation and choose the first
+                        	parameter of -c and -g (DEFAULT: 10)
+  		-o OUTPUT_NAME, --output_file_name OUTPUT_NAME
+                        	path to the output file in csv format (DEFAULT:
+                        	out.csv)
+  		-e EMOTION_IDS, --emotion_ids EMOTION_IDS
+                        	a list that contains emotion ids ranged from 0-39
+                        	(DEFAULT: 0). This can be a range expression, e.g.,
+                        	3-6,7,8,10-15
+  		-c C                SVM parameter (DEFAULT: 1). This can be a list
+                        	expression, e.g., 0.1,1,10,100
+  		-g GAMMA, --gamma GAMMA
+                        	RBF parameter (DEFAULT: 1/dimensions). This can be a
+                        	list expression, e.g., 0.1,1,10,100
+  		-s SCORE_DIR, --output_misc_dir SCORE_DIR
+                        	output intermediate data of each emotion in the
+                        	specified directory (DEFAULT: not output)
+  		-v, --verbose       show messages
+  		-d, --debug         show debug messages
+ 	```
+
+2. notes
+
+	* feature_list_file is in JSON format. Here is an example:
+		
+		```
+	 	[
+	    	{
+	        	"feature": "TFIDF_TSVD",
+	            "training": "adir/bdir/TFIDF_TSVD.train.npz",
+	            "testing": "cdir/ddir/TFIDF_TSVD.test.npz"
+	        },
+	        {
+	            "feature": "keyword",
+	            "training": "adir/bdir/keyword.train.npz",
+	            "testing": "cdir/ddir/keyword.test.npz"
+			}
+	    ]
+		```
+		
+    * Use example:
+
+    	```
+    	python batchSimpleTraining.py -k 10 -e 0-39 -o output.csv -c 1,10,100,1000 -v feature_list_ex.json
+    	python batchSimpleTraining.py -k 10 -e 0-39 -o output.csv -c 10,30,70,100,300,700,1000 -g 0.0001,0.0003,0.001,0.003,0.01,0.1 TFIDF_TSVD300.json
+
+		```
+		
 ## feelit/features.py
 
 1. Load features from files

diff --git a/batch/batchSimpleTraining.README b/batch/batchSimpleTraining.README
@@ -38,21 +38,22 @@ optional arguments:
   -d, --debug           show debug messages
 
 notes:
-    1. feature_list_file should look like this:
-        [
-            {
-                "feature": "TFIDF_TSVD",
-                "training": "adir/bdir/TFIDF_TSVD.train.npz",
-                "testing": "cdir/ddir/TFIDF_TSVD.test.npz"
-            },
-            {
-                "feature": "keyword",
-                "training": "adir/bdir/keyword.train.npz",
-                "testing": "cdir/ddir/keyword.test.npz"
-            }
-        ]
 
-    2. Use example:
+  1. feature_list_file is in JSON format. Here is an example:
+    [
+        {
+            "feature": "TFIDF_TSVD",
+            "training": "adir/bdir/TFIDF_TSVD.train.npz",
+            "testing": "cdir/ddir/TFIDF_TSVD.test.npz"
+        },
+        {
+            "feature": "keyword",
+            "training": "adir/bdir/keyword.train.npz",
+            "testing": "cdir/ddir/keyword.test.npz"
+        }
+    ]
+
+  2. Use example:
         python batchSimpleTraining.py -k 10 -e 0-39 -o output.csv -c 1,10,100,1000 -v feature_list_ex.json
         python batchSimpleTraining.py -k 10 -e 0-39 -o output.csv -c 10,30,70,100,300,700,1000 -g 0.0001,0.0003,0.001,0.003,0.01,0.1 TFIDF_TSVD300.json
 '''