An Innovative Method to analyze the Accuracy and Prediction rate of Handwritten Digit Recognition with Dimensionality Reduction Algorithm by comparing with Connectionist Temporal Classification

Recognizing and detecting the digits from the large unbiased data and to find the best accuracy and loss using machine learning algorithms such as Connectionist Temporal Classification (CTC) and innovative Dimensionality Reduction Algorithm (DRA).: Accuracy and Loss are performed with the MNIST dataset from the Keras library. The detection of digits is performed with the Digit's Dataset. The two groups Connectionist Temporal Classification (N=20) and Dimensionality Reduction Algorithms (N=20). A DRA is used for detecting the novel handwritten digits with the color represented to each digit. The accuracy is analyzed based on identifying the exact digit of 99.90% where the CTC has the accuracy of 90.84%. The two machine learning algorithms CNN and CTC are statistically satisfied with the independent sample T-Test (=.001) value (p!0.05) with a confidence level of 95%. Conclusion: Recognizing and detecting the handwritten digits significantly satisfied with better accuracy in DRA than in CTC.