Speech Recognition Using Matlab

Abstract Currently, technology is rapidly advancing to simplify people's daily lives. Our project aims to contribute to this by developing a voice recognition system that identifies the speaker and their speech, providing hands-free security that is resistant to hacking. The primary goal of this research is to create a more natural interface between humans and computers by enabling automatic speech recognition (ASR). Since people primarily communicate through speech, this technology can help remove communication barriers by allowing computers to understand human speech. To achieve this, we plan to design and implement a speech recognition system that accurately identifies a few simple commands. To recognize the speaker, we have used linear predictive parameters related to their vocal tract, along with hidden Markov models. Speech recognition is a widely used concept in security projects, as it helps confirm the identity of individuals automatically. In our project, we extract features using linear predictive coding from voice signals sampled directly from a microphone. We use Artificial Neural Networks (ANN) as the recognition method, with the codebook used as input feature vectors to the ANN. The back propagation algorithm of the ANN is used for training, and the training parameter is stored for future testing of the same speech sample. Additionally, diabetes is a prevalent disease affecting more than 246 million people worldwide, with this number expected to rise to around 380 million by 2025. It occurs when the body is unable to generate or respond correctly to insulin, which regulates glucose. This disease increases the risk of developing heart disease, kidney disease, blindness, nerve damage, and blood vessel damage. To diagnose diabetes, proper interpretation of the diabetes data is crucial, making it an important research issue. This study used a back propagation training algorithm for logistic regression classification with the diabetes dataset from Pima Indians. The dataset is a binary classification problem where the aim is to determine whether a patient has the disease based on several features.