NAMAQ - Arabic Handwriting Recognition Using Deep Learning, AI, and ML with Sentiment Analysis

The Arabic language is one of the delicate languages in terms of usage. Many characters in the language make it difficult for management to identify the script quickly. The holy book of the Muslims, the Quran, is written in Arabic. Several valuable resources from ancient and present times have helpful information on Arabic. The handwritten scripts vary from person to person, and it is not simple to understand and realize what is mentioned inside the scripts. In this study, the main emphasis is on recognizing Arabic handwritten words. The convolution neural network (CNN) is used, and feature extraction is done with the help of CNN for the Arabic text images uploaded into the system. With the help of machine learning classifiers, these handwritten scripts are recognized for similarity from the training data set. A recognition rate of up to 96.4% for 29 classes is achieved. The proposed model is tested using large datasets of characters and formations in the Arabic language. The identified features are further passed on to the machine learning classification model. Various researchers have worked on several models that yield variable rates of recognition. The highest rate is received at 96.3%. The next phase of the model proposed in this study focuses on the sentiment analysis of the scripts or texts classified in the recognition phase. The author's sentiment is identified for classifying the results.