人工神经网络在太阳能热水器市场预测中的应用

发布时间：2018-10-15 13:17

【摘要】：太阳能热水器是我国本土化并具有自主知识产权的产业。随着经济的发展，社会节能意识进一步提高，，大力促进了我国太阳能热水器产业的发展。“十二五规划”明确将太阳能产业作为我国战略新兴产业之一，出台大量的扶持政策。这些政策措施大力促进了太阳能热水器产业的蓬勃发展，行业的竞争也越来越激烈。而企业想在激烈的竞争中取胜，将必须争取以最合理的成本将产品交付给客户，这就要求企业要对市场的变化和业务本身的发展前景进行正确的评估和预测，这是现代企业成功的关键因素。预测是决策的前提，成功的决策离不开科学的预测。预测可以提高企业对不确定事件的反应能力，从而减少不利事件带来的损失，增加利用有利机会带来的收益。预测是针对市场变化规律可统计的范畴来进行的。传统市场预测模型多是利用时间序列内的历史需求数据来预测未来市场，且预测因子受个人经验判断的影响较大，使得预测技术的实际应用困难且预测精度较差。而人工神经网络因具有很强的自学习、自训练和非线性追溯能力，将有助于提高市场预测的精度和效率。本文通过与传统市场需求预测模型的比较，针对我国太阳能热水器市场需求的实际情况，设计研究出适用于太阳能热水器市场需求预测模型。首先，介绍了市场需求预测的相关理论和主要预测方法；其次，研究太阳能市场需求预测算法模型，比较Gompertz回归算法、指数平滑算法；再次，研究BP神经网络模型及其算法，应用该模型对太阳能热水器市场进行需求预测,并于Gompertz模型算法和指数平滑算法的预测结果相比较。
[Abstract]:Solar water heater is a native and independent intellectual property industry in China. With the development of economy, the consciousness of social energy saving has been further improved, which has greatly promoted the development of solar water heater industry in China. The 12th five-year Plan explicitly regards solar energy industry as one of our country's strategic emerging industries, introducing a large number of supporting policies. These policies and measures have greatly promoted the vigorous development of the solar water heater industry, and the competition in the industry is becoming more and more fierce. If the enterprise wants to win in the fierce competition, it must try to deliver the product to the customer at the most reasonable cost, which requires the enterprise to correctly evaluate and forecast the market change and the development prospect of the business itself. This is a key factor in the success of modern enterprises. Prediction is the premise of decision-making, and scientific prediction is indispensable to successful decision-making. Prediction can improve the ability of enterprises to respond to uncertain events, thus reducing the losses brought by adverse events and increasing the benefits of utilizing favorable opportunities. The forecast is made according to the statistical category of the law of market change. The traditional market forecasting models mostly use the historical demand data in time series to predict the future market, and the prediction factors are greatly influenced by personal experience judgment, which makes the practical application of forecasting technology difficult and the prediction accuracy is poor. Because of its strong self-learning, self-training and nonlinear traceability, artificial neural networks will help to improve the accuracy and efficiency of market forecasting. By comparing with the traditional market demand forecasting model, according to the actual situation of the solar water heater market demand in our country, this paper designs and studies the market demand forecasting model suitable for the solar water heater market. Firstly, it introduces the related theories and main forecasting methods of market demand forecasting. Secondly, it studies the algorithm model of solar energy market demand forecasting, compares Gompertz regression algorithm and exponential smoothing algorithm. Thirdly, it studies the BP neural network model and its algorithm. The model is used to predict the demand of solar water heater market, and compared with the prediction results of Gompertz model algorithm and exponential smoothing algorithm.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP183

【参考文献】