论文总字数：18457字

摘 要

数据的变化体现了经济时代的发展，从经济诞生以来它就与数据有着不可分割的联系，随着经济的不断鼎盛，股市也持续发展，投资者们能够从庞大的历史信息中，挖掘出许多重要的结论和有价值的相关信息，本文主要通过时间序列分析方法处理大量数据，深度学习利用数据，并建立恰当模型分析数据特征，从而实现银行行业的股票预测。

股票再为投资者带来高盈利的投资种类的同时,它也具有可能严重亏损的高风险,复杂的股票市场因其随着时间不断变动的价格常常让投资者十分困扰,一直以来大家都渴望提出一种理论来分析股价变动的原因,从而找到一种较为准确的方法指导投资者预测股票变化趋势从而规避投资风险,在股市取得更高的收益、降低损失。如何预测股票价格不仅是我国投资者关心的问题，更是世界范围内的难题，常因其不可限量的潜在经济利益而饱受全球大量的学者专家追捧。

对于投资者来说，最关心的问题自然就是怎样实现最大盈利或最小亏损。因为数据之间通常是息息相关的，数据的顺序和大小刻画了现实生活的持续变化，所以研究人员一般将一个时间序列比作一次随机过程进行分析。 所以我们可以认为，研究时间序列分析也可理解为观察数据在一段时间范围内呈现的统计规律性。本文将采用时间序列分析方法以工商银行2007-2019年季度收盘价为样本数据进行下一季度股价预测，模型预测结果5.7873与实际值5.51较为接近，表明本文采用的方法有效，对股票投资具有一定的指导作用。

关键词 ：股票分析 时间序列分析 股票预测 数据挖掘 ARIMA模型

Abstract

Data changes reflected the development of the economy, since the economy was born it has connected with data, with the constant height of economy, the stock market continues to develop, also can investors from huge historical information, dig out many important conclusions and valuable information, this paper mainly by the method of time series analysis processing of large amounts of data, the depth of the study using data, analyze data and establish the proper model characteristics, so as to realize the share of bank industry forecasts.

Stocks for investors with types of investments are extremely profitable at the same time, can also be a grievous loss of high risk, because of its complex stock market price often change over time for investors to very troubling, has long been everybody eager to put forward a theory to analyze the cause of the stock price changes,So, on the advice of our advisers, we will make our latest research available to scientists to avoid investment risk and achieve higher returns in the stock market.The prediction of stock price is not only a problem concerned by Chinese investors, but also a problem in the world.It is often sought after by a large number of scholars and experts around the world for its unlimited potential economic benefits.

Naturally, the main concern of investors is how to maximize profits or minimize losses.Because data are often closely related, researchers generally regard a time series as a random process. The order and size of the data in the time series also reflect the constant changes in the objective world.Therefore, we can think that the research significance of time series analysis lies in the observation of the statistical regularity of a certain time series in the process of long-term change.

In this paper, the time series analysis method will be used to forecast the stock price of icbc in the next quarter based on the sample data of the quarterly closing price of icbc from 2007 to 2019.The approximate value of 5.51 indicates that the method adopted in this paper is effective and has a certain guiding role in stock investment.

Keywords: stock analysis; time series analysis; stock prediction; data mining；ARIMA model;

目录

摘要............................................................Ⅰ

Abstract........................................................Ⅱ

1绪论...........................................................1

1.1研究背景与研究意义..........................................1

1.2国内外对股票趋势预测的研究..................................2

1.3股票时间序列分析预测原理....................................3

1.4本文的主要工作及其框架结构..................................3

2时间序列分析基本方法 .......................................5

2.1时间序列分析方法简介........................................5

2.2时间序列模型的构建..........................................6

2.3股票价格预测时间序列模型建立................................6

2.4时间序列模型的预处理........................................7

2.4.1差分运算................................................7

2.4.2平稳性检验..............................................8

2.4.3纯随机性检验............................................8

2.5时间序列基本模型............................................8

2.5.1 ARMA模型...............................................8

2.5.1.1模型定义............................................8

2.5.1.2模型性质............................................9

2.5.1.2建模步骤...........................................10

2.5.2 ARIMA模型.............................................10

2.5.2.1模型定义...........................................10

2.5.2.2建模步骤...........................................10

3基于时间序列的股票预测分析................................11

3.1股票时间序列分析...........................................11

3.1.1数据预处理.............................................11

3.1.2平稳化检验.............................................12

3.2股票时间序列模型建立.......................................13

3.2.1模型定阶...............................................13

3.2.2模型参数估计模型预测...................................15

3.2.3预测结果分析...........................................16

4 结束语...................................................... 17

4.1结论与建议.................................................17

参考文献...................................................... 18

致谢 ..........................................................19

请支付后下载全文，论文总字数：18457字