本文基于2007年2月7日至2019年10月30日的PTA期货数据，通过核密度估计方法对PTA期货价格走势进行核回归平滑，进而提取涨跌信号。并利用贝叶斯判别分析和回归分析模型建立涨跌走势与期货的各项技术指标的关联，从而对当日期货价格所处涨跌走势区间进行有效判断并对未来涨跌进行短期预测，防止“假涨（跌）”现象对判断期货价格走势造成干扰，帮助投资者更好的分析PTA期货价格的走势，并对期货价格涨跌拐点存在性的判断提供参考。

参考已有的股指期货价格走势分析方法，本文主要从中长期趋势（持续时长为一个月以上）与长期趋势（持续时长为一年以上）两个角度对PTA期货的价格走势信号进行提取，分别建立判别模型和回归预测模型，分析在不同周期中各技术指标对涨跌信号判断的影响以及模型的适用性。

Based on February 7, 2007 to October 30, 2019 data of PTA futures, kernel density estimation method is used for the smooth of PTA futures price movements with the extraction of signal. Moreover, Bayes discriminant analysis and regression analysis model are applied to establish the function which describe the relationship between moves of the PTA futures price with various technique related indexes ,to effectively judge the up or down trend range of futures price and predict future movements in short-term, prevent the phenomenon of "False Up (Down)" from interfering the judgement of futures prices’ trends of movements and the existence of a turning point.

Referring to existing methods of analyzing the price trend of stock index and futures, this paper mainly extracts the price trend signals of PTA futures of medium-term trend (lasting for more than one month) and long-term trend (lasting for more than one year), establishes discrimination model and regression prediction model respectively, and analyzes the influence of each technical index on rise and fall signal judgment in different periods with the applicability of the model.