<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[Hilda121的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/Hilda121</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; Hilda121]]></copyright><item><title><![CDATA[WSL成功安装]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/141636319</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/141636319</guid><author>Hilda121</author><pubDate>Wed, 28 Aug 2024 11:28:41 +0800</pubDate><description><![CDATA[成功安装WSL]]></description><category></category></item><item><title><![CDATA[阈值自回归模型（Threshold Autoregressive Model, TAR）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139582052</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139582052</guid><author>Hilda121</author><pubDate>Fri, 23 Aug 2024 10:08:42 +0800</pubDate><description><![CDATA[阈值自回归模型（Threshold Autoregressive Model, TAR）是一种用于处理非线性时间序列数据的统计模型。通过在不同状态下应用不同的线性自回归模型，TAR模型能够捕捉时间序列中的非线性特征]]></description><category></category></item><item><title><![CDATA[可解释人工智能（Explainable AI，XAI）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/140172245</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/140172245</guid><author>Hilda121</author><pubDate>Thu, 04 Jul 2024 10:16:40 +0800</pubDate><description><![CDATA[可解释人工智能（Explainable AI，XAI）是一组旨在让AI模型的决策过程变得透明和易于理解的技术和方法。随着AI系统在各个领域的广泛应用，了解其内部工作机制和决策依据变得尤为重要，尤其在需要高可信度和公平性的场景中，如医疗诊断、金融决策和法律判决等。]]></description><category></category></item><item><title><![CDATA[解释性AI，也称为XAI（Extended AI）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/140172077</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/140172077</guid><author>Hilda121</author><pubDate>Thu, 04 Jul 2024 10:12:36 +0800</pubDate><description><![CDATA[解释性AI，也称为XAI（Extended AI），是一种综合了多种AI技术的智能系统或工具，旨在帮助人们理解和解释数据、信息和知识，以辅助决策制定。XAI可以应用于各种领域，包括但不限于预测分析、风险评估、医疗保健、教育、人力资源、项目管理等。]]></description><category></category></item><item><title><![CDATA[多元线性回归、逻辑回归、Kmeans、dbscan、GBDT、XGBoost、Randomforest、Lightgbm、CNN、RNN、LSTM的特点、优势以及局限性]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139921832</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139921832</guid><author>Hilda121</author><pubDate>Mon, 24 Jun 2024 12:03:00 +0800</pubDate><description><![CDATA[多元线性回归、逻辑回归、Kmeans、dbscan、GBDT、XGBoost、Randomforest、Lightgbm、CNN、RNN、LSTM的特点、优势以及局限性]]></description><category></category></item><item><title><![CDATA[SAEs模型（Stacked Autoencoders，堆叠自编码器）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139917546</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139917546</guid><author>Hilda121</author><pubDate>Mon, 24 Jun 2024 10:11:10 +0800</pubDate><description><![CDATA[SAEs模型（Stacked Autoencoders，堆叠自编码器）是一种深度学习模型，用于无监督学习和特征提取。为了形象地解释这个模型，我们可以把它比喻成一个多层过滤器系统。]]></description><category></category></item><item><title><![CDATA[FISTA（Fast Iterative Shrinkage-Thresholding Algorithm）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139789889</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139789889</guid><author>Hilda121</author><pubDate>Wed, 19 Jun 2024 09:10:53 +0800</pubDate><description><![CDATA[FISTA（Fast Iterative Shrinkage-Thresholding Algorithm）是一种用于解决稀疏优化问题的高效算法，尤其适用于含有L1正则化项的凸优化问题。FISTA由Amir Beck和Marc Teboulle在2009年提出，是ISTA（Iterative Shrinkage-Thresholding Algorithm）的加速版本。]]></description><category></category></item><item><title><![CDATA[Nesterov加速梯度算法]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139789764</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139789764</guid><author>Hilda121</author><pubDate>Wed, 19 Jun 2024 09:04:52 +0800</pubDate><description><![CDATA[Nesterov加速梯度算法（Nesterov Accelerated Gradient，简称NAG）是优化领域中一种改进的梯度下降方法，由Yurii Nesterov于1983年提出。该算法的核心思想是在计算当前梯度之前，先使用上一次的动量更新进行预测，从而加速梯度下降的收敛速度。]]></description><category></category></item><item><title><![CDATA[AIC（Akaike信息准则）和BIC（贝叶斯信息准则）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139582209</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139582209</guid><author>Hilda121</author><pubDate>Mon, 10 Jun 2024 21:07:07 +0800</pubDate><description><![CDATA[AIC（Akaike信息准则）和BIC（贝叶斯信息准则）是用于模型选择的统计准则，它们帮助我们在多个候选模型中选择最佳模型。可以通过形象化的比喻和实例来解释这两个准则。]]></description><category></category></item><item><title><![CDATA[阈值自回归模型（Threshold Autoregressive Model, TAR）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139582066</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139582066</guid><author>Hilda121</author><pubDate>Mon, 10 Jun 2024 20:56:51 +0800</pubDate><description><![CDATA[阈值自回归模型（Threshold Autoregressive Model, TAR）是一种用于处理非线性时间序列数据的统计模型。通过在不同状态下应用不同的线性自回归模型，TAR模型能够捕捉时间序列中的非线性特征]]></description><category></category></item><item><title><![CDATA[岭回归（Ridge Regression）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139525621</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139525621</guid><author>Hilda121</author><pubDate>Fri, 07 Jun 2024 14:15:44 +0800</pubDate><description><![CDATA[岭回归（Ridge Regression），也叫Tikhonov正则化，是一种用于解决多重共线性问题的回归技术。通过引入L2正则化项，岭回归不仅可以防止模型过拟合，还能使回归系数变得稳定。]]></description><category></category></item><item><title><![CDATA[Lasso（Least Absolute Shrinkage and Selection Operator，最小绝对值收缩和选择算子）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139523544</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139523544</guid><author>Hilda121</author><pubDate>Fri, 07 Jun 2024 12:06:03 +0800</pubDate><description><![CDATA[Lasso（Least Absolute Shrinkage and Selection Operator，最小绝对值收缩和选择算子）是一种用于回归分析的技术，它通过引入L1正则化项来约束模型的系数，从而实现变量选择和模型简化。]]></description><category></category></item><item><title><![CDATA[Bi-LSTM（双向长短期记忆网络）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139501362</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139501362</guid><author>Hilda121</author><pubDate>Thu, 06 Jun 2024 15:14:12 +0800</pubDate><description><![CDATA[Bi-LSTM（双向长短期记忆网络）是一种高级的递归神经网络（RNN），能够同时考虑输入序列的前后文信息，从而更全面地理解文本的语义。为了形象地解释Bi-LSTM模型，可以用一个日常生活中的例子来帮助理解。]]></description><category></category></item><item><title><![CDATA[文本特征融合]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139500587</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139500587</guid><author>Hilda121</author><pubDate>Thu, 06 Jun 2024 14:52:03 +0800</pubDate><description><![CDATA[文本特征融合是一种将来自多个特征提取方法的特征组合在一起，以获得更全面和丰富的文本表示的方法。为了形象地解释文本特征融合，可以用一个日常生活中的例子来帮助理解。]]></description><category></category></item><item><title><![CDATA[门控机制（Gating Mechanism）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139500395</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139500395</guid><author>Hilda121</author><pubDate>Thu, 06 Jun 2024 14:46:25 +0800</pubDate><description><![CDATA[门控机制（Gating Mechanism）是一种在神经网络中特别用于控制信息流动的技术，广泛应用于递归神经网络（RNN）及其变体（如LSTM和GRU）。为了形象地解释门控机制模型，我们可以用一个日常生活中的例子来帮助理解。]]></description><category></category></item><item><title><![CDATA[多跳注意力机制（Multi-Hop Attention Mechanism）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139500152</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139500152</guid><author>Hilda121</author><pubDate>Thu, 06 Jun 2024 14:40:29 +0800</pubDate><description><![CDATA[多跳注意力机制（Multi-Hop Attention Mechanism）是一种高级的注意力机制，它通过多次关注输入中的不同部分来逐步提取更加细致和全面的语义信息。为了形象地解释这一机制，可以通过一个日常生活中的例子来理解。]]></description><category></category></item><item><title><![CDATA[凸松弛（Convex Relaxation）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139329488</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139329488</guid><author>Hilda121</author><pubDate>Thu, 30 May 2024 17:12:03 +0800</pubDate><description><![CDATA[凸松弛（Convex Relaxation）是一种优化技术，用于解决原本是非凸或组合优化问题的方法。]]></description><category></category></item><item><title><![CDATA[鲁棒性（Robustness）]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139329191</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139329191</guid><author>Hilda121</author><pubDate>Thu, 30 May 2024 16:58:17 +0800</pubDate><description><![CDATA[鲁棒性（Robustness）是指系统或模型在面对不确定性、噪声或扰动时，依然能够表现出良好性能的能力。]]></description><category></category></item><item><title><![CDATA[稀疏正则化器]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139329054</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139329054</guid><author>Hilda121</author><pubDate>Thu, 30 May 2024 16:55:18 +0800</pubDate><description><![CDATA[稀疏正则化器主要是指L1正则化。它的目的是使模型变得稀疏，即仅保留最重要的特征，并将不重要的特征的系数压缩为零。]]></description><category></category></item><item><title><![CDATA[L1和L2正则化]]></title><link>https://blog.youkuaiyun.com/Hilda121/article/details/139328817</link><guid>https://blog.youkuaiyun.com/Hilda121/article/details/139328817</guid><author>Hilda121</author><pubDate>Thu, 30 May 2024 16:50:56 +0800</pubDate><description><![CDATA[L1正则化和L2正则化是两种常见的正则化方法，用来防止机器学习模型的过拟合。我们可以用比喻的方式来形象地解释这两种正则化。]]></description><category></category></item></channel></rss>