decision
【五期李伟平】CCF-B（TFS'23）Consensus Reaching Process With Multiobjective Optimization for Large-Scale Group Decision Making With Cooperative Game
Peng Wu, Fengen Li, Jie Zhao, et al. Consensus Reaching Process With Multiobjective Optimization for Large-Scale Group Decision Making With Cooperativ ......
Markov decision process
https://medium.com/@hosamedwee/bellman-equation-1-understanding-the-recursive-nature-of-the-bellman-equation-in-mathematics-cd71bf14be1a https://mediu ......
POLIR-Management-TYPES of decisions
In a very simple sense, the problems managers encounter can be classified as: routine and familiar; new and unusual. In response, managers will use on ......
决策树(Decision Tree)
决策树是一种基于树结构的分类和回归模型，它通过对数据进行逐步的分解，从根节点开始，根据不同的特征进行分割，最终到达叶节点，叶节点对应一个预测结果。以下是决策树的基本概念和构建过程的详细解释： 决策树的基本概念： 节点（Node）： 根节点（Root Node）： 树的起始节点，包含整个数据集。 内部 ......
Management-Decision Making-{Rational,BoundedRational,Intuitive} D.M.
Management-Decision-{ Rational D.M.: Logical, Consistent and maximize value BoundedRational D.M.: "Good Enough" based on reality Intuitive D.M.: on th ......
Markov Decision Process Model Based on Value Iteration
Using the taxi example of OpenAI Gym to achieve and tune MDP model in Reinforcement Learning based on value iteration. ......
【AL&MT】Decision Tree
1 Introduction usual class in decision tree:ID3,C4.5,CART ID3:/Informattion Entropy,基于信息熵和信息增益 C4.5:/信息增益率,base on the ID3 CART:/基尼系数,using regress or ......
Paper Reading: NBDT: Neural-Backed Decision Trees
为了提高计算机视觉模型的可解释性，本文融合深度学习和决策树提出了神经支持决策树(NBDTs)。NBDT 使用一个可微的倾斜决策树取代了神经网络的最后一个线性层，和经典的决策树方法不同，NBDT 使用从模型参数派生的层次结构，不使用分层 softmax。NBDT 可以从任何现有的分类神经网络中创建，无... ......
《Decision Transformer: Reinforcement Learning via Sequence Modeling》论文学习
一、Introduction 先前的研究工作表明，Transformer可以对处于高维分布的语义概念进行大规模建模抽象，比较典型地体现如： 基于自然语言的零样本泛化（zero-shot generalization） 分布外图像生成（out-of-distribution image generat ......
Quantitative Approach of Management Science:(better decision making by using quantitative techniques)
Which is the use of **quantitative techniques to improve decision making**. Also known as _management science_. **Better decision making by using quan ......
Paper Reading: A three-way decision ensemble method for imbalanced data oversampling
针对 SMOTE 的缺点，本文提出了一种基于建设性覆盖算法(CCA)的三向决策抽样方法(CTD)。CTD 首先使用 CCA 构造不平衡数据的覆盖，然后选择少数样本的覆盖并根据覆盖的密度划分为三个区域。最后根据覆盖分布规律得到相应的阈值 α 和 β，选择关键样本进行SMOTE过采样。考虑到 CCA 随... ......
Paper Reading: Gradient Boosted Neural Decision Forest
[toc] Paper Reading 是从个人角度进行的一些总结分享，受到个人关注点的侧重和实力所限，可能有理解不到位的地方。具体的细节还需要以原文的内容为准，博客中的图表若未另外说明则均来自原文。 | 论文概况 | 详细 | | | | | 标题 | 《Gradient Boosted Neur ......
FIT5094 IT for Management Decision Making
FIT5094 IT for Management Decision MakingSemester 1, 2023Assignment 1 – Analysis of a Strategic DecisionFormat: Individual Report Weight: 25% of the m ......
Paper Reading: Interpretable Rule Discovery Through Bilevel Optimization of Split-Rules of Nonlinear Decision Trees
对于可解释的分类器本文限制为用简单的数学术语表示，使用非线性决策树(NLDT)将分类器表示为简单数学规则的集合。树的每个非叶结节点表示一个非线性数学规则，将给定条件节点中的数据集划分为两个不重叠的子集。通过限制每个条件节点上的分裂规则结构和决策树深度，保证了分类器的可解释性。在给定条件节点上的非线性... ......