diff --git a/.classpath b/.classpath
index b9832e4..fa09eea 100644
--- a/.classpath
+++ b/.classpath
@@ -3,8 +3,7 @@
-
-
+
diff --git a/Description.props b/Description.props
index 3482ded..e12ebdf 100644
--- a/Description.props
+++ b/Description.props
@@ -4,10 +4,10 @@
PackageName=NBSVM
# Version (required)
-Version=1.0.0
+Version=1.0.1
# Date
-Date=2016-03-01
+Date=2016-05-15
# Title (required)
Title=An NBSVM implementation capable of multiclass (one-vs-all) classification.
@@ -27,7 +27,7 @@ License=GPL 3.0
Description=This package contains a Java implementation of the NBSVM algorithm which was first presented as a binary classifier in "Baselines and Bigrams: Simple, Good Sentiment and Topic Classification," Sida Wang, Christopher D. Manning, in Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (ACL 2012), pp. 90-94, Jeju Island, South Korea (2012). This version extends the original algorithm to also support multiclass (one-vs-all) classification. The code relies on the LibLINEAR library and its Weka wrapper class. See the webpage for more information.
# Package URL for obtaining the package archive (required)
-PackageURL=https://github.com/vukbatanovic/NBSVM-Weka/releases/download/v1.0.0/NBSVM-Weka_1.0.0.zip
+PackageURL=https://github.com/vukbatanovic/NBSVM-Weka/releases/download/v1.0.1/NBSVM-Weka_1.0.1.zip
# URL for further information
URL=https://github.com/vukbatanovic/NBSVM-Weka/
@@ -36,4 +36,4 @@ URL=https://github.com/vukbatanovic/NBSVM-Weka/
Related=LibLINEAR
# Dependencies
-Depends=weka (>=3.7.12), LibLINEAR (>=1.9.5)
+Depends=weka (>=3.8), LibLINEAR (>=1.9.8)
diff --git a/README.md b/README.md
index 9189191..7d7ccb8 100644
--- a/README.md
+++ b/README.md
@@ -15,8 +15,8 @@ However, for the sake of consistency, the numbering of options used in LibLINEAR
## Installation
NBSVM-Weka can be installed as an unofficial plug-in module within Weka.
-To do so, download the [NBSVM-Weka package](https://github.com/vukbatanovic/NBSVM-Weka/releases/download/v1.0.0/NBSVM-Weka_1.0.0.zip).
-Open the Weka package manager (available in Weka >= 3.7) and use the "Unofficial - File/URL" option to select and install NBSVM-Weka.
+To do so, download the [NBSVM-Weka package](https://github.com/vukbatanovic/NBSVM-Weka/releases/download/v1.0.1/NBSVM-Weka_1.0.1.zip).
+Open the Weka package manager and use the "Unofficial - File/URL" option to select and install NBSVM-Weka.
After restarting Weka, the list of available classifiers (within the functions category) will contain the NBSVM option.
## Usage
@@ -25,7 +25,7 @@ Either way, the configuration options are very similar to the options of LibLINE
One difference is that NBSVM-Weka is not compatible with Support Vector Regression and Crammer and Singer's multiclass classification algorithm, making those SVMTypes disabled.
In addition, NBSVM-Weka allows the user to specify the Laplace smoothing parameter alpha (*default: 1.0*) and the interpolation parameter beta (*default: 0.25*).
-These parameters can be set in the command line options as -L (for Laplace smoothing) and -I (for interpolation).
+These parameters can be set in the command line options as -A (for Laplace smoothing) and -X (for interpolation).
## References
If you wish to use the NBSVM classifier in your paper or project, please cite the original paper:
@@ -34,7 +34,7 @@ If you wish to use the NBSVM classifier in your paper or project, please cite th
This Weka package was introduced in the following paper:
-**Reliable Baselines for Sentiment Analysis in Resource-Limited Languages: The Serbian Movie Review Dataset**, Vuk Batanović, Boško Nikolić, Milan Milosavljević, in Proceedings of the 10th International Conference on Language Resources and Evaluation (LREC 2016), Portorož, Slovenia. (2016)
+**[Reliable Baselines for Sentiment Analysis in Resource-Limited Languages: The Serbian Movie Review Dataset](http://www.lrec-conf.org/proceedings/lrec2016/pdf/284_Paper.pdf)**, Vuk Batanović, Boško Nikolić, Milan Milosavljević, in Proceedings of the 10th International Conference on Language Resources and Evaluation (LREC 2016), pp. 2688-2696, Portorož, Slovenia (2016).
## Additional Documentation
All methods contain extensive documentation and comments.
diff --git a/doc/allclasses-frame.html b/doc/allclasses-frame.html
index 181ffc2..a05ee89 100644
--- a/doc/allclasses-frame.html
+++ b/doc/allclasses-frame.html
@@ -2,10 +2,10 @@
-
+
All Classes
-
+
diff --git a/doc/allclasses-noframe.html b/doc/allclasses-noframe.html
index 0a35207..a54b515 100644
--- a/doc/allclasses-noframe.html
+++ b/doc/allclasses-noframe.html
@@ -2,10 +2,10 @@
-
+
All Classes
-
+
diff --git a/doc/constant-values.html b/doc/constant-values.html
index 0d66e0e..1696c42 100644
--- a/doc/constant-values.html
+++ b/doc/constant-values.html
@@ -2,10 +2,10 @@
-
+
Constant Field Values
-
+
@@ -122,7 +122,7 @@
An NBSVM implementation capable of multiclass (one-vs-all) classification. It relies on the LibLINEAR library and its Weka wrapper class.
- Sida Wang, Christopher D. Manning: Baselines and Bigrams: Simple, Good Sentiment and Topic Classification, in Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (ACL 2012), pp. 90–94, Jeju Island, South Korea. (2012) URL: nlp.stanford.edu/pubs/sidaw12_simple_sentiment.pdf
+ Sida Wang, Christopher D. Manning: Baselines and Bigrams: Simple, Good Sentiment and Topic Classification, in Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (ACL 2012), pp. 90–94, Jeju Island, South Korea (2012). URL: nlp.stanford.edu/pubs/sidaw12_simple_sentiment.pdf
@@ -167,6 +167,12 @@
Class NBSVM
-Z
Turn on normalization of input data (default: off)
+
+ -I <int>
+ The maximum number of iterations to perform.
+ (default 1000)
+
+
-P
Use probability estimation (default: off)
currently for L2-regularized logistic regression only!
@@ -181,10 +187,10 @@
Class NBSVM
-B <double>
Add Bias term with the given value if >= 0; if < 0, no bias term added (default: 1)
-
-L <double>
+
-A <double>
Set the value of the Laplace smoothing parameter alpha (default: 1.0)
-
-I <double>
+
-X <double>
Set the value of the interpolation parameter beta (default: 0.25)
-D
@@ -194,11 +200,11 @@
Class NBSVM
Version:
-
1.0.0
+
1.0.1
Author:
Vuk Batanović
See Also:
-
"Reliable Baselines for Sentiment Analysis in Resource-Limited Languages: The Serbian Movie Review Dataset", Vuk Batanović, Boško Nikolić, Milan Milosavljević, in Proceedings of the 10th International Conference on Language Resources and Evaluation (LREC 2016), Portorož, Slovenia. (2016)
+
Reliable Baselines for Sentiment Analysis in Resource-Limited Languages: The Serbian Movie Review Dataset, Vuk Batanović, Boško Nikolić, Milan Milosavljević, in Proceedings of the 10th International Conference on Language Resources and Evaluation (LREC 2016), pp. 2688-2696, Portorož, Slovenia (2016).
https://github.com/vukbatanovic/NBSVM-Weka
,
@@ -281,14 +287,14 @@
Field Summary
Fields inherited from class weka.classifiers.functions.LibLINEAR