Research On Classification And Prediction Of Game Evaluation Data

With the rapid development of current artificial intelligence, the application of artificial intelligence has provided people with a lot of convenience, has increased productivity, and has accelerated the paces of life and learning. We are improving a variety of artificial intelligence researches, such as pattern recognition, neural networks, evolutionary computation, and artificial immunology. These four issues can represent the four main methodologies of artificial intelligence, and with these four methods the artificial intelligence has been pushed to a wide development prospect. Along with the burgeoning of the big data, the neural network research is in a new climax period. In the field of evolutionary computation, the idea of bionics is used, and with the genetic algorithm the problem can be solved more precisely or optimized. The research on the artificial immune system is a new natural science, which has a wide prospect for the study, and there will certainly be widespread applications for human beings and society. There are both connection and difference among the above-mentioned four methodologies, which have their own unique ways and ideas to solve problems. But they can also be combined to achieve a better problem-solving purpose.The data classification&prediction problem is a very important issue and become more important in the era of big data. The data classification&prediction are necessary to solve many problems in the life, so the data classification&prediction are very valuable. For examples, the integration and analysis of Taobao’s commodity categories depend on the data classification, the data predication is useful in the study of Italian win, and the data classification method is also used in the stock prediction.At the present time, a large number of data are dealt with in most businesses, and the data have numerous information contents. With the diversification and high complexity of data types, the definition of big data arose. When the big data was put forward, it was considered as the determinants in all the fields. After the procedure of data collection, processing, feature extraction, algorithm research and data sharing, people can perceive the world in a deeper level and enjoy the progress of the society and the innovation of science and technology in a better way. At the present day, the cloud computing of the big data has become a very important field, and its data classification&prediction research becomes a very competitive issue. The classification and prediction of the game evaluation data are investigated in this paper, a variety of methods are used in the data analysis and data mining, in order to provide a firm foundation for the research of the big data.The data representation and processing are investigated first in this paper, and the data can be divided into three kinds. The first type is the digital data, the second is unstructured data, and the last one is the image data in the forms of pictures.For the text data, the theoretical study of the pattern recognition is made to make the classification and prediction of game data; the neural network is also investigated in the text classification of the game evaluation data; the immune genetic algorithm is designed to decrease the prediction error of the game evaluation text data; the support vector machine (SVM) is also investigated in the classification and prediction of the digital data and the unstructured data for the game. For the graphical data, the classification algorithms are improved in the feature extraction and classification prediction of the graphical data.The method for finding the weight of data elements with the SVM is proposed and verified by the experimental data comparisons. The algorithm for exacting the useful information from the big data samples is designed and verified useful with the experiments according to the classification accuracy. Based on the median voter law and the SVM, a new prediction method is put forward with high accuracy of data classification. The experimental results show that these new methods have obvious advantages in less running time and higher accuracy of classification and prediction.