《机器学习与模式识别的关系》由会员分享，可在线阅读，更多相关《机器学习与模式识别的关系（10页珍藏版）》请在金锄头文库上搜索。
1、题目：机器学习与模式识别学号：02115*姓名WWWMachine Learning and Pattern RecognitionAbstractRecently, machine learning has developed rapidly in information field. Also ,it has a close relationship with pattern recognition. Machining learning has been applied to pattern recognition successfully. Therefore , the paper
2、describes the basic characteristics of machine learning and pattern recognition, which includes the concepts ， development ， application and classification. It also provides an application perspective for understanding the concepts of machining and pattern recognition. Keywords: Machine Learning Pat
3、tern Recognition0. IntroductionMachine learning is one of the core problems of artificial intelligence research. Its application has been throughout all branches of artificial intelligence, such as expert systems, automated reasoning in the field of natural language understanding, pattern recognitio
4、n, computer vision, intelligent robotics. Just as its name implies, Machine learning is to let the computer to learn some way to improve its performance. Pattern recognition can be seen as something which can divide different objects into different categories. Humans can deepen their understanding o
5、f things through continuous learning, similarly the pattern recognition system based on simulating human intelligence also needs to improve its classification performance through machine learning algorithm improvements, so the contact between machine learning and pattern recognition is getting close
6、r and closer. This article will explain the basic concepts of machine learning and pattern recognition, pattern recognition analysis in several machine learning algorithms.1. Machine Learning1.1 The definition of machine learningCurrently , the accurate definition of machine learning : for certain a
7、ssignment T and performance metrics P, if a computer program to measure the performance of P and along with the experience of self-improvement on T, then we call the computer program is learning from experience E.1.2 The working mechanism of the machine learning systemThe environment provides certai
8、n information to the learning parts of the system, then the learning part uses this information to modify its knowledge base to enhance the performance of execution part; The execution do its work according the knowledge base, also bring back the acquired information to learning part. The process ca
9、n be seen as a certain process that the machine learning system acquire knowledge automatically with information which are provided by internal and external environment.1.3 The design of the machine learning systemThere are mainly two parts that need be taken into consideration when designing a perf
10、ect machine learning system : Model selection and design, Learning algorithm selection and design. Different models determine different objective functions and different learning mechanisms. The complexity and capacity of algorithm determine the capacity and efficiency of the learning system. Also t
11、he size of training samples and feature selection problem are the key factors which will constrain machine learning system performance.2. Machine learning algorithm in pattern recognitionPattern recognition means that we should analyze perception signal. It is a process of identification and interpr
12、etation. We can draw a picture toThe core issue of machine learning is searching problems. As for different application models, the researchers have designed some different searching algorithms. Currently in the field of pattern recognition, we often use genetic algorithms, neural networks, support
13、vector machines, k-nearest neighbor method and other machine learning algorithms.2.1 Genetic algorithmCharacteristic dimension is a major problem in machine learning, because the characteristics presented from certain model have different weights in reflecting the nature of things. But some showed n
14、o significant contribution to the catagories, even redundant, so the feature selectionprocess is very critical. Genetic algorithm can solve this problem to some extend as a optimization algorithm. Genetic algorithm not only can choose the feature that not only reflects the original information, but
15、also have a significant impact on the classification results.There are three kinds of operation in GA. Selection-reproduction, crossover, as well as mutation.We usually do as follows: Choose N chromosomes from population S in N separate times. The probability of one individual being chosen is P(xi).
16、 The computational formula of P(xi) :j =1There is a chance that the chromosomes of the two parents are copied unmodified as offspring, or randomly recombined (crossover) to form offspring. Also there is a chance that a gene of a child is changed randomly. Generally the chance of mutation is low.GA have four basic elements from the present: coding strategies;*!-厂 /*/*setting initial population; de
