Unique Digitized Activity Signature for Human Authentication

Activity recognition of human is one of the most important structure behind many implementations on devices like smart phones such as medical approaches, health tracing, context-awake mobile, human observe device, etc. This paper relates a robust device for activity recognition of human by smart phones. independent from other task, we explored the collection of instance selection and feature selection to decrease dimensionality of dataset in order to increase the presentation. Nowadays people give a little amount of time for any particular task. In this particular time nobody wants to waste his or her time for giving signature for any kind of security matching. In this paper, we are approaching a highly secured and time saving process for activity recognition called, `Activity Recognition by Digitized Method using Machine Learning Technique'. Our system consists of a smart phone which receives the details from its inbuilt sensors. Our concerned server received the sensors data from the smartphone. A unique map for an individual user is being generated by the server. The smartphone itself again sends the sensors details to the server and then again the server generates a unique map using those data and if the newly generated map is matched with previous one then the activity is recognized. This modern system is very useful and effective as there is less chances of fraud as the human own is the authentication signature.