https://www.luochang.ink/tags/rnn/ Recent content in RNN on Chang Luo Hugo zh-CN Sun, 12 May 2024 00:00:00 +0000 https://www.luochang.ink/posts/lstm_power_consumption/ Sun, 12 May 2024 00:00:00 +0000 https://www.luochang.ink/posts/lstm_power_consumption/ <blockquote> <p>长短期记忆网络（Long Short-Term Memory，LSTM）是一种 RNN 模型，常用于序列数据建模。尤其适合需要挖掘序列中长期依赖关系的问题。</p> </blockquote> <p>GitHub 项目地址：<a href="https://github.com/luochang212/lstm-model" target="_blank">luochang212/lstm-model</a></p> <h3 id="一keras-模型训练">一、keras 模型训练</h3> <ol> <li>数据导入</li> <li>数据预处理</li> <li>趋势可视化</li> <li>分割训练集和测试集</li> <li>定义 &amp; 训练网络</li> <li>预测</li> <li>预测整个序列</li> </ol> <br> <center> <button class="demo-btn" onclick="window_on('https://nbviewer.jupyter.org/github/luochang212/lstm-model/blob/main/1.keras模型训练.ipynb')">查看示例</button> </center> <br> <h3 id="二keras-模型推理">二、keras 模型推理</h3> <ol> <li>数据预处理</li> <li>用训练好的 keras 模型做预测</li> <li>预测下一个值</li> <li>预测下一个序列</li> </ol> <br> <center> <button class="demo-btn" onclick="window_on('https://nbviewer.jupyter.org/github/luochang212/lstm-model/blob/main/2.keras模型推理.ipynb')">查看示例</button> </center> <br> <h3 id="三pytorch-模型训练">三、PyTorch 模型训练</h3> <blockquote> <p>施工中</p> </blockquote> <p>参考资料：</p> <ul> <li> <p><a href="https://www.kaggle.com/datasets/sagnikseal/household-power-consumption" target="_blank">Household_Power_consumption</a> (dataset)</p> </li> <li> <p><a href="https://www.kaggle.com/code/akdagmelih/ocean-wave-prediction-with-lstm/notebook" target="_blank">Ocean Wave Prediction with LSTM</a></p> </li> <li> <p><a href="https://www.kaggle.com/code/amirrezaeian/time-series-data-analysis-using-lstm-tutorial/notebook" target="_blank">Time-series data analysis using LSTM (Tutorial)</a></p> </li> <li> <p><a href="https://docs.kanaries.net/pygwalker#a-python-library-for-exploratory-data-analysis-with-visualization---pygwalkers" target="_blank">PyGWalker</a></p> </li> <li> <p><a href="https://github.com/luochang212/rnn-note" target="_blank">luochang212/rnn-note</a></p> </li> </ul> <div id="mini-overlay" onclick="overlay_off()"></div> <div id="mini-window"><iframe id="mini-iframe" frameBorder="0"></iframe></div> <button id="btn-close" onclick="overlay_off()">×</button> <script src="https://www.luochang.ink/mini-viewer/overlay.js"></script> <link rel="stylesheet" href="https://www.luochang.ink/mini-viewer/style.css">