Abstract: With the popularity of the various security system, identify the authentication of an exact user has been a key challenge for each system. The voiceprint recognition technology has some merit over another security system with flexible, economical terms. Therefore using voice as a key security tool will be useful for the ecosystem. In this system, we are going to use some machine learning concept SVM (Support Vector Machine). The SVC will be useful to differentiate the dataset and find out the actual required result so that user can get authentication to access the machine (Files, Folder, Documents), to make the system fully workable it is supported by Python on the frontend, with the help of the python SVC implement of it is successful and from the backend part, we are going to use IBM Watson which will provide actual power to the system. Using the IBM server APIs would be extra key to success to make the system more efficiently get the result as solving machine learning process on the client is hectic and hard to perform, that’s why we are using the server of IBM which provides some free tools which are easily accessible and taking advantage of these tools we can easily achieve the required security and result for the system.

Keywords: Machine Learning, Python, IBM Watson Text to Speech, Fuzzy Wuzzy, Skitlearn, Support Vector Machine (SVM), Mel-Frequency Cepstral Coefficients.

| DOI: 10.17148/IARJSET.2021.8603