DDAG-based SVM multi-class classification active learning algorithm

Abstract

The invention discloses a DDAG-based SVM multi-class classification active learning algorithm. The active learning through is adopted in the multi-class SVM learning process, the defects that according to the traditional 'one-to-many' or 'one to one' multi-class classification method, large amount of indivisible points or decision blind areas are provided during sample testing are utilized, the active learning algorithm based on the 'o-v-o' classification decision blind areas is provided, samples with the 'highest uncertainty' of the indivisible points or decision blind areas corresponding to a current learning device are selected actively, according to the limitation caused by the single active learning strategy during active learning, a multi-strategy integration active learning method on the basis of posterior probability and similarity measurement uncertainty is provided in the multi-class SVM learning process, the two active learning methods are combined effectively, the sample labeling load is reduced in the multi-class SVM learning process, the learning sample labeling cost is reduced, and an SVM classification device with the best performance can be obtained through the least labeled sample training.

Description

technical field [0001] The invention relates to an algorithm, in particular to a DDAG-based SVM multiclass classification active learning algorithm, belonging to the technical field of machine learning algorithms. Background technique [0002] SVM (Support Vector Machines, Support Vector Machines) is a new pattern recognition method developed on the basis of the VC dimension theory of statistical learning theory and the principle of structural risk minimization. It can seek the best compromise between the complexity of the model (i.e., the learning accuracy for a specific training sample) and the learning ability (i.e., the ability to identify any sample without error) based on limited sample information, in order to obtain the best generalization ability. It largely solves the problems of model selection and over-learning, nonlinear and dimension disasters, local minimum points and other problems existing in traditional pattern recognition technology. Many unique advantag...

AI Technical Summary

Problems solved by technology

[0020] 1) When the number of categories is large, the number of training samples of a certain category will be much less than the sum of the number of training samples of other categories, and the unbalanced distribution of training samples will affect the classification accuracy;

[0021] 2) When the number of training samples n and the number of categories K are large, the speed of training and testing classification is very slow;

[0022] 3) When the test sample x does not belong to any of the K classes or has more than one class with the largest decision function value, there is a problem of rejection or decision blindness

[0028] 3) The number of classifiers increases sharply with the number of classes K, resulting in slow classification decisions;

[0029] 4) When testing classification, if there are two or more categories with the same number of votes, the system will not be able to determine which category they belong to, that is, the method has the problem of rejecting points or blind spots in decision-making

[0042] 1) There is a phenomenon of "error accumulation" from top to bottom;

[0043] 2) The number of classifiers increases sharply with the number of classes K, and there are problems of large amount of calculation and long training time

[0051] For example, the active learning method based on error reduction needs to search the entire sample space before selecting samples. For a large amount of unlabeled sample sets, this sample selection strategy directly calculates the classifier on the test data set after adding samples. Classification error, the complexity of its calculation is quite high, it is not feasible in practice

[0052] (2) It is easy to collect meaningless samples

For example, active learning based on uncertainty sampling may sample outliers, and it is difficult to distinguish samples with a large amount of information from abnormal samples.

[0054] (3) Influence of error propagation

[0055] If the learner trained in the initial stage of active learning is inaccurate, the sample selected during the active learning process may not be the "most beneficial" sample for the learner training, which will affect the performance of active learning

[0056] (4) Active learning does not have incremental learning ability

[0057] Active learning is an iterative process. Every time a new sample is added, the classifier must be retrained. However, the general active learning method does not have the ability to learn incrementally, and one of the problems faced by active learning is that repeated iterative operations bring High time complexity and space complexity

[0058] (5) Need more manual participation and feedback

[0074] However, the current above-mentioned methods simply apply active learning strategies to multi-class classification problems, and do not consider the role of the "disadvantage" characteristics of the specific multi-class classification methods in the algorithm in active learning, that is, Figure 4 The role of process P2 in active learning. At the same time, these algorithms have certain application limitations. For example, it is difficult to determine the query function to measure "the most beneficial" or "most useful", and the calculation is complicated. In general, in the active learning process P3, only Using a single active learning method, instead of using the complementary effects of active learning strategies to minimize the cost of learning sample labels, while improving the performance of the classifier

Images