中文摘要：

针对当前手写数字识别正确率较低这一不足,提出了一种主成分分析（PCA）和粒子群算法优化支持向量机（PSO-SVM）的手写数字识别方法。首先,利用PCA降低输入数据的维数,然后把降维的数据作为SVM的输入,用PSO不断优化SVM中的核函数参数g和惩罚因子c,以提高分类精度。实验结果表明：同传统的SVM、GA-SVM、网格搜索算法、卷积神经网络（CNN）相比,PSO-SVM方法分类方法具有最高的识别准确率且运算效率也较高,达98.2%,性能上优于其他几种分类算法。

英文摘要：

In this paper, a new method of handwritten numeral recognition based on principal component analysis （PCA） and particle swarm optimization （PSO-SVM） is proposed for the problem of low accuracy of handwritten digit recognition. Firstly, the dimension of the input data is reduced by PCA, then the dimension reduction data is used as the input of SVM, and the kernel function parameter g and the penalty factor c in SVM are optimized by PSO to improve the classification accuracy. The experimental results show that SVM and GA-SVM, with the traditional grid search algorithm, convolutional neural network （CNN） compared with the classification method of PSO-SVM method and it has higher recognition accuracy rate and the operation efficiency is the highest, reached 95.2%, and the performance is better than other types of classification algorithms.