基于类邻域字典的线性回归文本分类

文本表示的高维性会增加文本分类时的计算复杂度.针对该问题,构建基于类邻域字典的线性回归分类模型.采用K近邻方法构造各类别的类邻域字典,根据对测试样本的不同表示,分别提出基于级联类邻域字典和基于类邻域字典的线性回归分类算法.此外,为缓解噪声数据对分类性能的影响,通过度量测试样本与各个类别之间的相关度裁剪噪声类数据.实验结...

Linear Regression Text Classification Based on Class-wise Nearest Neighbor Dictionary

In text classification, the high dimensionality of text representation increases the computational complexity. To address the problem, a Linear Regression Classification(LRC) model is constructed based on neighborhood dictionary. The K-Nearest Neighbor(KNN) method is used to construct the neighbor dictionary for each class, and the LRC algorithms based on the concatenate class-wise nearest neighbor dictionary and the class-wise nearest neighbor dictionary are proposed separately according to the different representations of the test sample. In addition, the correlation between the sample and the classes is measured to clip the noise data, alleviating the impact of noise data on classification performance. The experimental results show that the proposed model provides high classification accuracy and calculation efficiency for long texts and short texts. For those texts with multiple classes, the strategy of noise class clipping also enables it to display excellent classification performance.