دانلود کتاب بصری سازی و تجسم ترکیبی با استفاده از یادگیری ماشین با Mathematica – پایتون

This second edition of Hybrid Imaging and Visualization book adds four new topics: Fisher discriminant, which is a linear discriminant that can provide an optimal separation of objects (Sect. 2.6), and converting time series into images thereby making it possible to employ convolution neural network to classify time series effectively (Sect. 3.6). Optimizing hyperparameters is an important task in machine learning and mostly, stochastic global methods are used. Among others, the fancy Black Hole algorithm is introduced and compared with other more usual methods (Chap. 6), and ChatGPT a novel and in the last two years very popular Generative AI technology is introduced and its ability is illustrated (Chap. 7). Fisher discriminant presented in Sect. 2.6 can be used as a supervised learning classifier. Given labeled data, the classifier can find a set of weights to draw a decision boundary, classifying the data. Fisher’s linear discriminant attempts to find the vector that maximizes the separation between classes of the projected data similar to the support vector machine (SVM) method. Often, in the real world, a linear discriminant is not complex enough to separate datasets effectively. To deal with nonlinear separations, one should employ Fisher discriminant with different kernels. Different examples are given in this section to demonstrate this technique. Converting time series into image discussed in Sect. 3.6 is useful in case of classification as well as clustering of time series since convolutional neural network (CNN) can handle images very effectively. In this section, the Gramian Angular Field method is introduced and demonstrated using a toy example. Optimizing hyperparameters presented in Chap. 6 is very important for efficient use of Machine Learning (ML) algorithms, although it requires heavy computing load. The problem comes from two sources (a) the basic algorithm should be repeated many times, and (b), in most cases, there are more local optimums, consequently global optimization methods should be employed. The efficiency of different global optimization techniques is demonstrated using the case of an image classification problem. Finally, the ChatGPT, which is nowadays the most popular AI topic is introduced in Chap. 7, where its principle and application areas are discussed and demonstrated. Perhaps the most perplexing issue is how one employs it for writing computer codes and getting an explanation of its detailed meaning. One should however keep in mind that the more one is experienced in computer programming and able to define the characteristics of the needed code, the more efficient the ChatGPT results will be.