The digital revolution converts artificial intelligent in too many subsets like Machine Learning, Deep Learning etc. In some other words we can say AI is a science of machines which trains them to perform human tasks while Machine learning is make a system able to learn automatically. In this paper we try to recapitalize the practices of machine learning for a business environment. We also discuss some important points have to follow by a developer of generic algorithm for ML.

A machine can be trained and obtained data by the technique used by machine leaning on experiences and examples Traditionally we have to write long codes for that but in Machine learning their algorithm feeds to base algorithm. These MLA or machine builds the logic based on the feed data instead of writing the lengthy code.

software defined network (SDN) and virtualization network function (NFV) are cooperatively perceived as the most encouraging bearing for adaptable programmability of organization control capabilities and conventions with dynamic use of organization assets. SDN gives the deliberation of organization assets over obvious application programming interface to accomplish fundamental topology independent numerous inhabitant networks with required Quos and service level agreement (SLAs). NFV worldview sends network capabilities as programming occasions, specifically, VNFs on item equipment utilizing virtualization procedures. Along these lines, virtual IP capabilities, for example, load adjusting, steering, and sending or firewall, can work as VNF in a cloud with a positive result in network execution. In this paper, we meant to accomplish traffic load adjusting by utilizing a virtual SDN (SDN) regulator as NFV. With SON, when there is lopsided and expanded load, optional SDN regulators can be added to share this heap. The need of optional not entirely settled and a duplicate SON with the very same setups as unique SON is made, which works precisely and shares traffic load offsetting errands with a unique SON regulator. Both SON regulators are autonomously put in the cloud with straight forwardness guaranteeing that each client in the organization knows all about the presence of the recently made auxiliary SON regulator. We tentatively approved the heap adjusting in Fat-Tree geography involving two SDN regulators in a Mininet emulator. The outcomes showed half improvement in normal burden, 41% improvement in normal deferral, and significant enhancements concerning ping reaction, data transmission use, and throughput of the framework.

This paper presents the development and performance analysis of Adaptive Neuro-Fuzzy Inference System (ANFIS) based MPPT controller for a DC to DC converter. The proposed system consists of 2.0 kW PV array, DC to DC boost converter and load. The proposed algorithm has advantages of neural and fuzzy networks. To enhance of converter performance, Adaptive Neuro-Fuzzy Inference System (ANFIS) based MPPT controller is used. In order to demonstrate the proposed ANFIS controller abilities to follow the reference voltage and current, its performance is simulated and compared with Artificial Intelligence Technique based MPPT controller. Simulation results show that for a wide range of input irradiance, Adaptive Neuro-Fuzzy Inference System (ANFIS) based MPPT controller shows improved performance than the Artificial Intelligence Technique based MPPT controller with at various operating conditions.