Abstract -- When a customer buys and sell cars, it’s a great challenge to fix the accurate price. ANN-Artificial Neural Network, a branch of AI-Artificial Intelligence, are commonly used for such calculations. This project is mapped out into two different artificial neural networks to perform regression task. They are car price prediction and its tested using data from a car sales website. The libraries are imported from pendas, a speedy numpty, MPLX, matplotlib and Seabourn it used for data plotting, visualizations & numerical analysis. Penda is mainly used for data frame manipulations to import data into data frame Cardiff. A special type of encoding used to deal with all the special characters. So, it uses cycler learn, a package used extensively for machine learning and deep learning training in general and also in this project, The min max scalar and Sai kat learn is used to perform normalizations. This project includes optimizers called Adam Optimizers to compile the model and gradient descent or the kind of back prorogation. So, it has some customer data, such as customer name, customer email, the country, the customer, gender, age, annual salary, credit card debt and the net worth simply. These are all the parameters that predict how much customers are willing to pay for a new vehicle for a new car. With actual predictions a kind of targeted marketing is achieved to that specific customer. It shows cars within that price range to customers, because the model knows that's what a customer can afford in a way. So, this project aims to show cheaper cars or more expensive cars to customers. The objective is to show the price range of cars to perform Target Marketing on the basis of Car purchase amount and annual salary of customer. It really enhances marketing in its kind of larger scheme of Things. A success rate of 92% will be obtained and more dataset is required for accuracy.

Abstract -- Generally the renewable sources are affected by the environmental variations. The change of the environmental conditions has a great impact in the energy generation and load utilization. Uncertainty associated with load profile and intermittent renewable have created a larger gap in marinating the balance between loads and sources. To design a smart grid that contains multiple renewable sources .To design an optimal energy management scheme for a smart grid .To perform a energy management scheme based on the optimal energy allocation which contains renewable sources To manage the smart grid’s energy distribution which suffers from the energy uncertainties.

Abstract -- In brain, the growth of abnormal cells is called brain tumour some of which may lead to cancer. To detect the brain tumour by usual method is Magnetic Resonance Imaging (MRI) scans. About the abnormal tissue growth in the brain is identified from the MRI images information. The detection of brain tumour is done by applying Machine Learning and Deep Learning algorithms. When these algorithms are applied on the MRI images the prediction of brain tumour is done very fast and a higher accuracy helps in providing the treatment to the patients. This prediction also helps the radiologist in making quick decisions. In this proposed work to detecting the presence of brain tumour and their performance is analysed by Convolution Neural Network (CNN).

Abstract -- In this study, by inserting an additional switched network into the traditional buck-boost converter, a new transformer less buck-boost converter is proposed. Using the knowledge of open loop converter behavior, a closed loop converter is designed. DC-DC converter will be designed for specific line and load conditions. But in practice there is deviation of the circuit operation from the desired nominal behavior due to changes in the source, load and circuit parameters. So we need to design a proper controller or compensator to overcome this situation of the circuit operation. This work presents PI controller designed such that any input variations produces a constant output voltage.

Abstract -- Face-recognition technologies are becoming increasingly popular in today's society. Facial recognition technology is widely used in security systems. It's crucial to consider the face-recognition system's ability to survive an unauthorized person's attack, though. The use of fake images and videos can compromise face recognition systems. The aim of this paper was to develop Face liveness detection. Face liveness detection's primary goal is to differentiate between a real and a fake face and also to detect whether it’s live or not. Face anti-spoofing is essential to prevent face recognition systems from a security breach. Although face anti-spoofing detection methods have been proposed so far, the problem is still unsolved due to the difficulty on the design of features and methods for spoof attacks. According to a 2018 report by the Center for Applied Internet Data Analysis (CAIDA), there are close to 30,000 spoofing attacks per day. Face liveness detection is a preprocessing step in face recognition for avoiding face spoofing attacks. To overcome this problem Convolutional Neural Networks (CNN) based Face liveness detection method is introduced. This proposed system classifies the real or fake face Anti-spoofing technologies come in help in these situations to avoid these assaults. Real and fake faces are distinguished using the obtained face embedding’s, which are then concatenated and delivered via Softmax layer for classification (Softmax layer as a Classifier). The proposed method for detecting spoofs has a test accuracy of 89% and a specificity of 90%.