ANN Based POS Tagging For Nepali Text

International Journal on Natural Language Computing (IJNLC) Vol.7, No.3, June 2018
DOI: 10.5121/ijnlc.2018.7302 13
ANN Based POS Tagging For Nepali Text
Archit Yajnik
Department of Mathematics, Sikkim Manipal University, Sikkim, India
ABSTRACT
The article presents Part of Speech Tagging for Nepali Text using three techniques of Artificial Neural
networks. The novel algorithm for POS tagging is introduced .Features are extracted from the marginal
probability of Hidden Markov Model. The extracted features are supplied to 3 different ANN architectures
viz. Radial Basis Function (RBF) network, General Regression Neural Networks (GRNN) and Feed
forward Neural network as an input vector for each word. Two different Annotated Tagged sets are
constructed for training and testing purpose. Results are compared using all the 3 techniques and applied
on both the sets. GRNN based POS tagging technique is found better as it produces 100% and 98.32%
accuracies for both training and testing sets respectively.
KEY WORDS
Radial Basis Function, General Regression Neural Networks, Feed forward neural network, Hidden
Markov Model, POS Tagging
1. INTRODUCTION
Natural Language Processing (NLP) is a diversified field of computational linguistics which in
high demand for the researchersworld wide due to its large number of applications like language
translation, Parsing, POS tagging etc. Among those POS tagging is one of the core part of NLP
which is being used in other applications of computational linguistics. There are various
techniques available in the literature for POS tagging like HMM based Viterbi algorithm, SVM
etc.Artificial neural networks plays a vital role in various fields like medical imaging, image
recognition is covered in [1, 2, 3] and since last one decade it becomes popular in the field of
Computational linguistics also. Due to the computational complexities sometimes it is not
preferred for the big data analysis. General Regression Neural Network which is based on
Probabilistic neural networks is one type of supervised neural network is computationally less
expensive as compared to standard algorithms viz. Back propagation, Radial basis function,
support vector machine etc is exhibited in [4]. That is the reason GRNN is considered for the Past
of speech Tagging experiment for Nepali text in this article. Following sentence illustrates
annotated tagged Nepali sentence
(My Name is Archit)
Several statistical based methods have been implemented for POS tagging [5] as far as Indian
languages are concern. Nepali is widely spoken languages in Sikkim and neighbouring countries

14
like Nepal , Bhutan etc. The use of ANN architecture is seldom for tagging [6]. To develop a
parser and Morphological analyser for the natural languages POS tagging plays a pivotalrole.
This article presents a neural network architecture based on the Statistical learning theory
described in [4]. This neural network is usually much faster to train than the traditional multilayer
perceptron network. This article is divided into five sections. After this introduction, the second
section briefs the traditional ANN architectures viz Feed forward MLP, Radial basis function
networks (RBF), General Regression Neural Network (GRNN). Experimental set up is
highlighted in the third section followed by Result and discussion in the 4th
section. The article is
concluded in the fifth section.
2. ANN ARCHITECTURES
Single layer feedforward Multilayer Perceptron is based on backpropagation algorithm using
Sigmoidal transfer function is successfully employed in various multiclass classification
problems. Due to the existence of Centres, the kernel based technique RBF is widely used for the
classification. The networks can be trained for centres as well as synaptic weights. The detail
information about these networks is available in [4].
GRNN is also a kernel based technique in which Mahalanobis distance is calculated of each
pattern with the corresponding centre. The detailed information about Probabilistic and General
Regression Neural Networks is available in [4]. GRNN can briefly be introduced for the training
set, { }N
iii y 1
),( =
x . To estimate the joint probability distribution for vectors x and y say ),( yf Y xX,
and therefore )(xXf , we may use a nonparametric estimator known as the Parzen – Rosenblatt
density estimator. Basic to the formulation of this estimator is a kernel, denoted by K(x), which
has properties similar to those associated with a probability density function:
Assuming that x1,x2, … ,xN are independent vectors and identically distributed (each of the
random variables has the same probability distribution as the others), we may formally define the
Parzen – Rosenblatt density estimate of )(xXf as
where the smoothing parameter h is a positive number called bandwidth or simply width; h
controls the size of the kernel. Applying the same estimator on ),( yf Y xX, , the approximated
value for the given vector x is given by
If we take Gaussian kernel i.e.
2
)( X
x −
= eK , we obtain,

15
where )()(2
ii xxxx −−= T
iD and σ is the standard deviation. )(ˆ xf can be visualized as a
weighted average of all observed values yi , where each observed value is weighted exponentially
according to its Euclidean distance from x. The theory of General Regression Neural Networks
discussed above is pertaining to only neuron in the output layer. The same technique can be
applied for the multiple neurons in the output layer also. Therefore the technique can be
generalized as shown below:
Let wij be the target output corresponding to input training vector xi and jth
output node out of the
total p. Again let Ci be the centres chosen from the random vector x. Then
Here n be the number of patterns in the training set. The estimate yj can be visualized as a
weighted average of all the observed values, wij, where each observed value is weighted
exponentially according to its Euclidean distance from input vector x and n is the number of
patterns available in the input space
3. EXPERIMENTAL PROCEDURE
The survey of Part of Speech Tagging for Indian languages is covered by Antony P J (2011) in
[7]. The details of the tags used for the experiment is available in [8, 9]. Out of which 42100
samples (patterns) are used for training and the remaining 2500 samples for testing. The database
is distributed in to n = 41 tags. Network architecture consists of 41 x 3 = 123 input neurons,
42100 hidden neurons which plays a role of centres Ci (i = 1, 2, …,42100) shown in (4) and 41
neurons in output layer.
Transition and Emission probability matrices are constructed for both the sets
viz. training and testing. Transition matrix demonstrates the probability of occurrence of one tag
(state) after another tag (state) hence becomes a square matrix 41 x 41. Whereas the emission

16
matrix is the matrix of probability distribution of each Nepali word is allotted the respective tag
hence it is of the size n x m (number of Nepali words) . In order to fetch the features for ith
word
say , the ith
row, ith
column of the transition matrix and ith
row of the emission matrix are
combined hence becomes 41 x 3 = 123 features for each word. Therefore the ANN architectures
consists of 123 input neurons All the patterns (or Nepali words) are used as a centre. Euclidean
distance is calculated between patterns and centres. Training set consists of 5373 words hence the
same number of hidden neurons are incorporated in ANN architectures.
As there are 41 tags, 41 output neurons constitute the output layer of the network. For instance if
the word belongs to NN (common noun) category which is the first tag of the tag set then the first
neuron has a value 1 and all others are 0.
3.1 FEED FORWARD NEURAL NETWORK
The MLP network is trained with 123 input neurons, 30 hidden neurons and 41 output neurons
using sigmoidal transfer function till the error goal of computed output and original output
reachesuptoExp(-16). The code is implemented in Matlab. The network is trained up to 800
epochs.
3.2 RADIAL BASIS FUNCTION NETWORK
The RBF network is trained with 123 input neurons, hidden neurons are same as the number of
patterns and 41 output neurons using sigmoidal transfer function till the error goal of computed
output and original output reachesuptoExp(-16). The code is implemented in Matlab. The
network is trained up to 800 epochs.
3.3 GENERAL REGRESSION NETWORK
The synaptic weights are computed for GRN network as shown in section 2. There are 123 input
neurons, hidden neurons are same as the number of patterns and 41 output neurons using
Gaussian transfer function The code is implemented Java.
4 RESULT ANALYSIS
The training database contains 42100 words whereas testing set consists of 6000 words.Both
database does not contain the words with multiple tags i.e. same nepali word with more than one
different tags. The performance of all the three networks are depicted in table 1. Experiments
demonstrate that all the architectures are better able to identify the suitable tags as far as the
training set is concern but in the case of testing set except GRNN none other network performs
efficiently. Table 1 exhibits the fact that 5899 out of 6000 testing samples are tagged properly
where in other cases tagging accuracy is extremely poor.

17
Table: 1 Performance analysis of ANN techniques
As demonstrated the output of GRNN is portrayed on the test input shown below. Only the first
word नानीनानीनानीनानी(a child) is erroneously tagged PP (pronoun) but its actual tag is NN (common noun). It
has happened because the frequency of the word नानीनानीनानीनानीis 1 in the training set and the frequency of
PP is highest that is 37 among all the tags.
5 CONCLUSIONS
GRNN achieves better output accuracy (98.32%) in testing set as compared to other two ANN
techniques viz RBF and Feed Forward Neural Network as depicted in table 1. Words of testing
set with less frequency of occurrence may be 1 or 2 in training set are failed to get an appropriate
tag in the case of GRNN architecture. It concludes that GRNN technique which does not require
any kind of training and yields the output only based on the distance provides much better
accuracy as compared to the other ANN architectures which require training.

18
ACKNOWLEDGEMENTS
The author acknowledges Department of Science and Technology, Government of India for
financial support vide Reference no SR/CSRI/28/2015 under Cognitive Science Research
Initiative (CSRI) to carry out this work.
REFERENCES
[1] Richard O Duda and Peter E Hart, “Pattern Classification”, 2006, Wiley-Interscience, New York,
USA.
[2] S. Rama Mohan, ArchitYajnik: “Gujarati Numeral Recognition Using Wavelets and Neural
Network” Proceedings of Indian International Conference on Artificial Intelligence 2005, pp. 397-
406.
[3] ArchitYajnik, S. Rama Mohan, “Identification of Gujarati characters using wavelets and neural
networks” Artificial Intelligence and Soft Computing 2006, ACTA Press, pp. 150-155.
[4] Simon Haykin, “Neural Networks A Comprehensive Foundation” Second Edition, Prentice Hall
International, Inc., New Jersey, 1999.
[5] Prajadip Sinha et al. 2015.Enhancing the Performance of Part of Speech tagging of Nepali language
through Hybrid approach, 5(5) International Journal of Emerging Technology and Advanced
Engineering.
[6] Tej Bahadur Shai et al. 2013. Support Vector Machines based Part of Speech Tagging for Nepali
Text, Vol: 70-No. 24 International Journal of Computer Applications.
[7] Antony P J et al. 2011.Parts of Speech Tagging for Indian Languages: A Literature Survey,
International Journal of Computer Applications (0975-8887), 34(8).
[8] http://www.lancaster.ac.uk/staff/hardiea /nepali/ postag.php
[9] http://www.panl10n.net /english/ Outputs%20 Phase %202/CCs/Nepal/MPP/ Papers/2008/ Report
%20 on %20 Nepali %20 Computational %20 Grammar.pdf .
[10] ArchitYajnik, “Part of Speech Tagging Using Statistical Approach for Nepali Text”, International
Journal of Computer, Electrical, Automation, Control and Information Engineering Vol:11, No:1,
2017, pp. 76-79.

ANN Based POS Tagging For Nepali Text

More Related Content

What's hot (17)

Similar to ANN Based POS Tagging For Nepali Text (20)

Recently uploaded (20)

ANN Based POS Tagging For Nepali Text