天津农学院学报 ›› 2021, Vol. 28 ›› Issue (2): 72-78.doi: 10.19640/j.cnki.jtau.2021.02.015

• 研究与简报 • 上一篇    下一篇

基于机器学习的梭鲈鱼生长发育模型

典彩华1, 田云臣1,2,*   

  1. 1.天津农学院 计算机与信息工程学院,天津 300392;
    2.天津市水产生态与养殖鱼点实验室,天津 300392
  • 收稿日期:2020-09-14 发布日期:2021-07-26
  • 通讯作者: 田云臣(1967—),男,教授,硕士,主要从事农业信息技术研究工作。E-mail:tianyunchen@tjau.edu.cn。
  • 作者简介:典彩华(1983—),女,硕士在读,研究方向为水产信息化。E-mail:longago172@sohu.com。
  • 基金资助:
    财政部和农业农村部:国家现代农业产业技术体系(CARS-47); 天津市海水养殖现代农业产业技术体系(ITTMRS2021012)

Growth and development model of pike perch based on machine learning methods

Dian Caihua1, Tian Yunchen1,2,*   

  1. 1. College of Computer and Information Engineering, Tianjin Agricultural University, Tianjin 300392, China;
    2. Tianjin Key Lab of Aqua-ecology and Aquaculture, Tianjin 300392, China
  • Received:2020-09-14 Published:2021-07-26

摘要: 基于大数据、人工智能建立鱼类生长发育模型以实现水产养殖精细管理已成为关注的热点。本文介绍了梭鲈鱼养殖数据的来源以及采集、定性特征的处理等数据预处理方法,然后介绍了利用因子分析和主成分分析两种方法进行降维处理的步骤,最后详细介绍了机器学习数据集划分方法和运用BP神经网络、支持向量机、XGBoost 3种方法建立梭鲈鱼生长发育模型并对模型进行验证的方法和过程。结果表明,基于随机划分数据集、采用BP神经网络和主成分分析方法建立的模型对梭鲈鱼生长发育的预测效果最好。

关键词: 梭鲈鱼, 生长发育模型, 机器学习, 数据处理, 影响因子

Abstract: Establishing fish growth and development models based on big data and artificial intelligence technology to achieve fine management has become a hotspot in the aquaculture industry. This article introduced the data preprocessing methods such as the source and collection of pike perch farming data and qualitative feature processing, then the steps of dimensionality reduction using factor analysis and principal component analysis, and finally the method of dividing the machine learning data set and the steps of using BP neural network, support vector machine, XGBoost to establish the growth model of pikeperch. The analysis of the established models showed that the models based on randomly divided data sets, BP neural network and principal component analysis are the best predictors for the growth and development of pikeperch.

Key words: Pike perch, growth model, machine learning, data preprocessing, impact factor

中图分类号: