Authors :
Md Ziaul Haque; Mohd Omar
Volume/Issue :
Volume 7 - 2022, Issue 7 - July
Google Scholar :
https://bit.ly/3IIfn9N
Scribd :
https://bit.ly/3AfSKZq
DOI :
https://doi.org/10.5281/zenodo.6982395
Abstract :
Hindi is a national language of India spoken in
many states in our countries, like Bihar, Uttar Pradesh,
Madhya Pradesh, Jharkhand, and Delhi. The Hindi
language is 3rd most popular language globally, which is
the script of Devanagari. It consists of 36 primary
alphabets and ten digits. We present sophisticated
handwritten Hindi character recognition (2HCR) using
machine learning techniques to implement Hindi
characters and digits. A dataset consists of Ninety-Two
Thousand images of 46 different types of characters and
digits in the Hindi language segmented from handwritten
documents. Nowadays, it has become easy totrain data
because of the availability of various algorithms and
methodology. We have used many classification
algorithms for implementing and improving accuracy.
Classification algorithms are Linear-Regression (LR),
Logistic-Regression (LGR), Support-VectorMachine (SVM), Random-Forest (RF), and Naïve-Bayes
(NB) to classify the model and improve the accuracy.
Handwritten Character Recognition, the area for
research is still an active platform because of individuals’
different human writing styles, shapes, and sizes. Also, it
is used in many applications such asreading license plate
numbers, document reading, cheque numbers, postcodes
on envelopes, verification of signatures, etc. This system,
that we have developed, designed, and implement, has
been done using python programming. After completing,
we analyzed the performance and accuracy of thesystem.
Keywords :
Machine Learning, Python, 2HCR, OCR, Hindi Character, Devanagari, LR, LGR, SVM, RF, NB.
Hindi is a national language of India spoken in
many states in our countries, like Bihar, Uttar Pradesh,
Madhya Pradesh, Jharkhand, and Delhi. The Hindi
language is 3rd most popular language globally, which is
the script of Devanagari. It consists of 36 primary
alphabets and ten digits. We present sophisticated
handwritten Hindi character recognition (2HCR) using
machine learning techniques to implement Hindi
characters and digits. A dataset consists of Ninety-Two
Thousand images of 46 different types of characters and
digits in the Hindi language segmented from handwritten
documents. Nowadays, it has become easy totrain data
because of the availability of various algorithms and
methodology. We have used many classification
algorithms for implementing and improving accuracy.
Classification algorithms are Linear-Regression (LR),
Logistic-Regression (LGR), Support-VectorMachine (SVM), Random-Forest (RF), and Naïve-Bayes
(NB) to classify the model and improve the accuracy.
Handwritten Character Recognition, the area for
research is still an active platform because of individuals’
different human writing styles, shapes, and sizes. Also, it
is used in many applications such asreading license plate
numbers, document reading, cheque numbers, postcodes
on envelopes, verification of signatures, etc. This system,
that we have developed, designed, and implement, has
been done using python programming. After completing,
we analyzed the performance and accuracy of thesystem.
Keywords :
Machine Learning, Python, 2HCR, OCR, Hindi Character, Devanagari, LR, LGR, SVM, RF, NB.
