java递归神经网络大棚温室短序列预测+源代码
毕业论文关键词：自动化温室；递归神经网络；长短记忆网络；时间序列
Based on the recursive neural network prediction of short series for greenhouse
Abstract： With the progress of the society development, the traditional mode of agricultural production can't meet the needs of people. New type of facilities agricultural—greenhouses, has a merit of that is not limited to time or space. It is widely used in agricultural production in various kinds of environments. Compared with traditional greenhouse, automated greenhouse has incomparable advantages. Tranditional automated greenhouse controls the environmental factors by complex linear function model. Neural network estimates all of the environmental factors, then realizes the control. So neural network is easier and faster than the tranditional way. Recursive neural network has excellent effect in time series prediction in the numerous neural network. While the ordinary recursive neural network is not good in learning effect for long-term dependence on information, LSTM - an improved recursion network model, can effectively solve the problem. So in case of the automated greenhouse under the complex environment factors, adopt LSTM can have better prediction effect.This paper mainly includes three aspects:(1) Firstly, it’s the characteristics of the data processing and different approaches to process data.(2) Secondly, constructs the network model and test different parameters in the model based on the learning and calculating to the environmental factors of the greenhouse. And makes a contrast adjustment to achieve the best learning effect. (3) Finally, writes a graphical interface, makes it easy for operation by a visual interface.
Key words: Automated Greenhouse ;Recurrent Neural Network ;Long Short-Term Memory ;Time Series
目 录
摘要： 1
关键词： 1
Abstract： 1
Key words: 1
1 绪论 2
1.1自动化温室研究背景及现状 2
1.2递归神经网络研究背景及现状 2
1.3 研究内容和技术路线 3
1.3.1研究内容 3
1.3.2 技术路线 4
2 递归神经网络结构及相关函数介绍 4
2.1 RNN 4
2.2 LSTM 7
2.3 激活函数 9
2.3.1 Sigmoid 函数 9
2.3.1 Tanh 函数 10
2.3.2 ReLu 函数 11
2.4 优化函数 11
2.4.1 SGD 12
2.4.2 RMSprop 12
2.4.3 Adam 12