Smart personal assistant using machine learning

Assisting users in performing their tasks is the main goal of todays personal assistant applications. Many such applications are being developed, which are capable to discover the user's habits, abilities, preferences, and goals, even more accurately and predicting the user's actions in advance and perform them without user's interaction. The assistant agent has to continuously improve its behavior based on previous experiences. Improvements are achieved in personal assistant applications by learning mechanism. Agents are capable of accessing information from databases to guide people through different tasks, deploying a learning mechanism to acquire new information on user behaviour. Also the resources has to be utilised in highly efficient way leading to less power consumption. In this paper we have proposed a machine learning approach for learning mechanism of personal assistant agent. For better utilization of smartphones battery, the interrupt based broadcast receiver approach has been followed.