HINDIA: a deep-learning-based model for spell-checking of Hindi language

Abstract

The spelling error is a mistake occurred while typing the text document. The applications like search engines, information retrieval, emails, etc., require user typing. In such applications, good spell-checker is essential to rectify the misspelling. Spell-checkers for western languages like English are very powerful and can handle any type of spelling errors, whereas in the case of Indian languages like Hindi, Urdu, Bengali, Kannada, Assamese, etc., the available spell-checkers are very basic ones. These spell-checkers are developed using traditional methods like statistical methods and rule-based methods. This article presents a novel model HINDIA to handle the spelling errors of the Hindi language, one of the most spoken languages in India. It utilizes a deep-learning method for spelling error detection and correction. The proposed spell-checking model works in two phases. In the first phase model identifies the erroneous words in the input sample and in the second phase it replaces the wrong words with the most probable correct words. Model HINDIA is developed using the attention-based encoder–decoder bidirectional recurrent neural network (BiRNN) which uses long short-term memory cells. Several modifications in the BiRNN have been made and network is fine-tuned to process the spelling errors of Hindi language. It uses publicly available dataset ‘monolingual corpus’ developed by IIT Mumbai for training and testing. The performance of the proposed model is evaluated in two scenarios. In the first scenario where the testing dataset is generated using split function. HINDIA performs significantly well with precision 0.86, recall 0.72, f-measure 0.78 and accuracy 0.80. Further, in the second scenario, where a dataset is manually generated its performance is fairly good with precision 0.81, recall 0.72, f-measure 0.76 and accuracy 0.74. Model HINDIA gives better performance than the deep-learning-based Malayalam spell-checker and some other deep-learning-based correction models present in the literature.

Appendices

Appendix A: List of variables used in the article and their definitions

Sr. no.	Variables	Definition
1.	X  = (X1, X2, X3, …, Xn)	Fixed size input vector
2.	H  =  (H0, H1, H2, …, Hm)	Hidden layers
3.	Y  =  (Y0, Y1, Y2,…,Yk)	Output symbols
4.	X t	Input to the network at time t
5.	H t	Number of Hidden layers at time t
6.	Y t	Output symbol at time t
7.	V t	Current word symbol at time t
8.	W	Weight matrix of the input vector
9	U	Recurrent weight matrix
10.	B	Bias of the network
11.	V	Weight of the hidden layer
12.	f t	Activation function of forget gate at time t
13.	W f	It is the weight matrix of forget gate
14.	B f	Bias of the forget gate
15.	Σ	Sigmoid function
16	i t	Activation function of input gate at time t
17.	$\hat{C}_{t}$	Cell input activation vector
18.	W c	Weight matrix of the candidate
19.	b c	Bias of the candidate cell
20.	C t	Cell state at time t
21.	O t	Output at time t
22.	W o	Weight matrix of output gate

Appendix B: List of abbreviations

Sr. no.	Abbreviation	Full-form
1.	RNN	Recurrent neural network
2	BiRNN	Bidirectional recurrent neural network
3.	LSTM	Long short-term memory
4.	AI	Artificial intelligence
5.	DL	Deep-learning
6.	NLP	Natural language processing
7.	SMS	Short message service
8.	POS	Part-of-speech
9.	HMM	Hidden Markov model
10.	REDM	Reverse edit distance model
11.	FSA	Finite state automata
12.	FSR	Finite state representation
13.	SCRNN	Semi character recurrent neural network
14.	FFNN	Feed-forward neural network
15.	BPTT	Backpropagation through time
16.	En-De RNN	Encoder–decoder recurrent neural network
17.	PoO	Probability of occurrence
18.	FAQ	Frequently asked question
19.	CBOW	Continuous bag of word