









摘" 要: 隨著工業生產的復雜性和規模的增加,準確的工業預測模型對于提高生產效率和降低成本至關重要。為此,提出一種基于tSNE?LSTM算法的工業預測模型來預測工業生產過程的溫度。將t?SNE應用于數據降維和特征提取,然后使用LSTM進行序列建模和預測。該模型結合了t?SNE降維和LSTM循環神經網絡的優勢,能夠有效地捕捉時間序列數據的非線性動態特征,并進行準確預測。在實際工業數據集上的實驗驗證結果表明,該模型在工業預測任務中具有較高的準確性和魯棒性。
關鍵詞: 工業預測; 溫度預測; t?SNE; LSTM; 時間序列數據; 非線性動態特征
中圖分類號: TN911?34; TP18" " " " " " " " " " "文獻標識碼: A" " " " " " " " " " " 文章編號: 1004?373X(2024)12?0081?05
Industrial prediction model based on tSNE?LSTM algorithm
TAN Jiansuo1, WU Xinghua2, XU Wenguang1, YANG Kaiming1, XING Xiangyun1, WANG Hongliang2
(1. Tin Chemical Company Limited of Yunnan Tin Co., Ltd., Mengzi 661019, China;
2. Faculty of Civil Aviation and Aeronautics, Kunming University of Science and Technology, Kunming 650500, China)
Abstract: As the increases of complexity and scale of industrial production, accurate industrial forecasting models are essential to improve production efficiency and reduce costs. On this basis, an industrial prediction model based on tSNE?LSTM algorithm is proposed to predict the temperature of industrial production process. t?SNE is used for the data dimensionality reduction and feature extraction, and then LSTM is used for the sequence modeling and prediction. The advantages of t?SNE dimensionality reduction and LSTM recurrent neural network are combined in this model, so that the nonlinear dynamic features of time series data can be captured" effectively and be predicted accurately. By the experimental verification on actual industrial datasets, the results show that the model has high accuracy and robustness in industrial prediction tasks.
Keywords: industrial forecasting; temperature predition; t?SNE; LSTM; time series data; nonlinear dynamic features
0" 引" 言
隨著工業生產的智能化和自動化程度的提高,以及工業領域中大規模、高維時間序列數據的快速增長,如何準確預測和分析這些數據成為了一個重要的挑戰。針對這個問題,近年來出現了基于機器學習的預測方法,其中基于深度學習的方法已經在工業預測中得到了廣泛應用。
范國棟等人為了提高工業生產效率和安全性,提出了基于XGBoost算法構建的工業機械設備故障預測模型,和經過決策樹訓練出來的工業機械設備故障類型預測模型,具有較高的準確性[1]。蔣建香等人以工業用戶作為城市用電主體,提出一種基于VMD?GRU?EC的工業用戶短期負荷預測方法。該方法具有由負荷組成結構復雜、易受生產計劃等因素影響而產生沖擊性負荷的特點。其次,在采用傳統的VMD?GRU預測值的基礎上,進一步考慮增加基于門控循環網絡的誤差修正方法對初始預測值進行修正,從而更好地適應工業負荷的多變性和突變性[2]。……