基于电子病历的乳腺癌群组与治疗方案可视分析

乳腺癌是当前最常见的恶性肿瘤之一，其电子病历数据可用于挖掘隐含规律，对治疗与预后分析有重要意义。通过与乳腺科医生合作，选择合适的预测模型和可视化方法，搭建了一个基于电子病历的乳腺癌群组和治疗方案可视分析系统。首先，对具有高维属性的病人进行降维和聚类处理，形成病人群组，并采用南丁格尔图、词云和时间轴可视化方法，直观展示病人群组间特征的差异；然后，用支持向量机（support vector machine，SVM）模型预测治疗方案，用平行坐标、矩阵热力图和分类图分别展示属性相关性、训练后的特征权重和预测结果；最后，用真实案例验证了系统在群组分析、治疗方案及病人属性关联分析中的有效性，从而较好地帮助医生选择合适的治疗方案。

Digital calligraphy generation method

It is of important significance for the spread of calligraphy art by using digital technology to display traditional calligraphy art on electronic screens. In this paper we present an approach of digital calligraphy generation to promote the art of calligraphy. Firstly, it employs Kalman filtering and quadratic spline interpolation respectively to smooth the trajectory of the brush on the screen according to the speed of the brush movement. Secondly, the calligraphy rendering with high degree of simulation is generated by adopting a brushstroke morphological restoration method which is based on the real writing art. Lastly, in order to obtain a better effect of Half-Dry Stroke in calligraphy, a rule-based approach is applied to modify the calligraphy handwriting.