BP算法的改进及其在股票价格预测中的应用
ImProvedBPAlgorithmand115APPlicationinPre山etionofStockPrice
-
摘要: 为加速BP算法的收敛,提出了一种物理意义明确、体现人脑优选本质的新的激励函数,通过动态调整 此激励函数的参数并结合已有的一些BP改进算法,用之进行股票价格的预测,得到了满意的结果。同时,股票 价格的可预测性也从另一角度证明了我国不成熟股票市场的非有效性。只要预测模型选取恰当,可获得超过市 场平均盈利水平的收益。Abstract: InordertospeedtheeonvergeneeofBPalgorithm,thisPaPerPutsfo哪ardanew stimulationfunetion,whiehhasanexPlieitphysiealmeaningand15abletorefleettheoptimization esseneeofhumanbeings,brains.ThroughdynamieallyadjustingtheParametersofthestimulation funetionandeolnbiningotherimprovedBPalgorithms,theauthorsusethestimulationlunetionto predietthePrieeofastoekandgetasatisfaeto巧result.ontheotherhand,thepredietabilityofstoek prieesalsoProvesthenoneffeetiveoftheimmaturestoekmarketsofChinafromanotherangle.Aslong a,aproperpredietionmod6115ehosen,peopleeangetmoreprofitsthanaveragemarketearnings.
-
Key words:
- neuralnetworks /
- Predietions /
- stimulationfunetion /
- stoekprice
点击查看大图
计量
- 文章访问数: 1414
- HTML全文浏览量: 52
- PDF下载量: 359
- 被引次数: 0