中文摘要：

癫痫是神经科常见疾病，具有反复发作和难以治愈的特征。对癫痫发作的预测可以使医护人员或患者提前采取有效措施来降低或避免癫痫发作所带来的损伤。从20世纪90年代至今，癫痫预测算法得到了长足发展，同时也面临着如算法精确度、数据处理方式、预测能力有限等方面的问题。本文对上述问题进行了整理阐述，并对癫痫预测算法的发展趋势做了进一步分析。

英文摘要：

Epilepsy is a kind of common disease in Neurology. It has the characteristics of repeated attack and intractability. Seizure prediction can help nursing staff and patients themselves adopt effective methods in advance to lower or prevent patients from possible injuries from seizures. From 1990s to date, development in Seizure Prediction Algorithm （SPA） has gained great progress totally. Meanwhile, there are some problems in this area, including degree of accuracy, way of date processing and prediction ability, etc. This paper aims to introduce the problems mentioned above and analyze the future development in SPA.