- 博客(72)
- 资源 (2)
- 收藏
- 关注
原创 Agile Development
Agile development: Make you won’t do extra workSprintProduct Backlog:User StoryTimeboxingSprint Planning MeetingSprint Review Meeting: deliverablesRetrospective meeting:Daily meeting:KISS: Keep it simple, stupid!Treat warnings as errors
2023-09-19 22:43:35
290
原创 HTML-CSS-js教程
双标签单标签html5的DOCTYPE声明用于定义文档的头部。文档的头部包含了各种属性和信息,包括文档的标题,在Web中的位置以及和其他文档的关系等。绝大多数文档头部包含的数据都不会真正作为内容显示给读者定义文档的主体,包含文档的所有内容,在页面中显示出来,用户可以直接看到的内容定义文档的标题,可以显示在浏览器的标签栏或状态栏上。标签是标签中唯一必须要求包含的东西,就是说head一定要写title,用来描述HTML网页文档的属性,关键词等。例如。
2023-02-10 11:49:23
642
1
原创 词嵌入和交叉熵|Word Embedding and Cross Entropy
词嵌入,交叉熵损失函数,Word Embedding,Cross Entropy,Pytorch
2022-10-28 12:32:04
725
原创 Transformer和BERT学习笔记|Notes for BERT and Transformer
对每一个下游任务,构造一个跟这个任务相关的神经网络,将于训练好的表示,如词嵌入,作为一个额外的特征,和本来任务的输入一起输入到模型中。我希望这些特征已经有了比较好的表示,所以导致模型训练起来比较容易这两个途径都是使用一个相同的目标函数:都是使用语言模型,并且是单向的 (unidirectional language models)Bert 的改进:Masked Language model。
2022-10-12 20:32:43
571
原创 代理标签的半监督学习算法|Proxy-label Approaches for Semi-supervised Learning
Tri-training with disagreement 在原算法基础上做了个小调整(算法2Co-training中的第8行),要求对于模型。应该拒绝预测,也就是disagreement。置信度较高的样本,第三个模型。
2022-10-11 11:48:37
442
原创 自然语言处理教程-注意力模型|Natural Language Processing with Attention Models
NLP,Transformer,Bert
2022-10-07 06:15:39
477
原创 应用深度学习课程笔记|Review of 4995 Applied Deep Learning
Deep Dream里面能够改变的参数不是filter,而是input image,通过调整input image,来使得output image as high as possible (excites the layer)大部分的参数来自于最后一层dense layer,通过global average pooling可以极大减少参数数量。
2022-09-26 11:02:47
864
原创 数据科学AWS实践1-AutoML|Analyze Datasets and Train ML models using AutoML
AWS, AutoML, Sagemaker, S3, Autopilot
2022-08-19 01:24:58
501
1
原创 前端网络框架Bootstrap教程|Front-End Web UI Frameworks and Tools: Bootstrap4
前端,Front end,Bootstrap,JavaScript
2022-07-27 03:23:25
318
原创 吴恩达深度学习笔记3|Coursera Deep learning Notes - Structuring Machine Learning Projects
Deep learning, 吴恩达深度学习教程笔记
2022-07-25 00:34:43
426
原创 可解释机器学习- InterpretML的使用|interpretable machine learning- InterpretML tutorial
可解释机器学习,InterpretML
2022-07-15 11:49:13
1917
原创 可解释机器学习- LIME模型讲解|interpretable machine learning-LIME
可解释机器学习,interpretable machine learning,LIME,模型讲解
2022-07-15 11:47:08
4253
原创 Python中的多进程多线程|Multiprocessing and Multithreading in Python
进程,线程,多进程,多线程,一些concurrent package的实例
2022-07-12 21:48:02
816
原创 吴恩达深度学习笔记及作业答案|Coursera Deep learning Notes - Improving Deep Neural Networks
Deep learning, dropout (inverted dropout), Early stopping, normalization, Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization
2022-07-12 04:33:08
2961
原创 Review of 4705 NLP
ContentsLecture 12: Lexical Semantics (part I) - Word Representations and Word Embeddings.Lecture 13: Machine Learning: Linear and Log-Linear ModelsLecture 14: Machine Learning: Feed-forward Neural Networks, Autoencoders/embeddings, Dense networksAutoencod
2022-05-10 12:45:59
205
原创 Review of 4121 Computer System for Data Science
ContentsComputer systems and performance rules of thumbData centersDatabasesStorage and distributed file systemsDistributed systemsMapreduceDistributed analytics and streamingCachingAdditional informationComputer systems and performance rules of thumbLa
2022-04-24 08:57:40
651
原创 Review of 4721 Machine Learning
ContentsMultivariate GaussiansReview: Variance and covarianceMultivariate GaussiansPCA/ SVDMultivariate GaussiansReview: Variance and covarianceCovariance matrix: X⃗=(X1,X2,…,Xd)\vec{X} = (X_1, X_2, \dots, X_d)X=(X1,X2,…,Xd), cov(X⃗)cov(\vec{X})cov
2022-04-21 09:38:22
1797
原创 Review of Linear Algebra
ContentsTranspose and Inverse(AB)−1=B−1A−1(AB)^{-1} = B^{-1} A^{-1} (AB)−1=B−1A−1(AT)−1=(A−1)T(A^T)^{-1} = (A^{-1})^{T}(AT)−1=(A−1)TAA−1=I, (A−1)T(AT)=IAA^{-1} = I,\ (A^{-1})^T(A^T) = I AA−1=I, (A−1)T(AT)=IPermutation matrixP−1=PT P^{-1} =
2022-04-07 12:01:17
344
原创 Review - 5703 Statistical Inference and Modeling
ContentLecture 1 IntroLecture 2 EstimationEstimation:Two estimation methodsOptimality in estimationLecture 3 Confidence intervals and hypothesis testingLecture 1 IntroLLN (Law of large numbers)CLT (Central Limit Theorem)CMT (Continuous Mapping Theorem
2022-02-08 04:06:28
875
原创 Review of 4246 Algorithms for Data Science
ContentsImportant algorithmsNoteLecture1: Insertion sort, efficient algorithmLecture2: Merge sortLecture3: Binary search, quicksortLecture5: Graphs, Breadth-First Search (BFS)Lecture6: Depth-first search, topological sortingLecture7&8: Strongly connect
2021-12-28 00:40:38
497
原创 Cracking Metric/Business Case/Product Sense Problems
Table of contents1. What are product sense/business case/metric problems2. 3 categories of questions and frameworks2-1 Diagnose a problem2-2 Measure success2-3 Launch or not1. What are product sense/business case/metric problems2. 3 categories of ques
2021-12-04 06:21:20
135
原创 Product Case Interviews
Table of contentsTipsIt's okay to make mistakes.Interact with the interviewerAlways clarify questionsAsk time to structure your communicationRed flagYou have no idea at allYou have to many ideasFollowing Frameworks BlindlyNot able to defend your selfFacebo
2021-12-04 01:48:04
406
原创 Summary of Statistics for Interview
Table of ContentsP_valuePower of a test / statistical powerP_valuehttps://en.wikipedia.org/wiki/P-valueIn null hypothesis significance testing, the p-value is the probability of obtaining test results at least as extreme as the results actually observ
2021-11-15 07:40:41
548
原创 SQL刷题总结 SQL Leetcode Review
Table of contentsnull, '', 0null, ‘’, 0这三个值是不同的。对于0,直接用=检查即可。对于“”,也可以直接用=检查。但是需要注意的是,0 = ‘’是成立的。所以需要单独删除0的情况。eg: 176with temp as ( select distinct salary, dense_rank() over(order by salary desc) as r from Employee)select IF( (select t
2021-11-09 04:11:19
771
1
原创 AB Testing Review
Cracking A/B Testing Problems in DS interviewhttps://towardsdatascience.com/7-a-b-testing-questions-and-answers-in-data-science-interviews-eee6428a8b63power = 1 - (Type 2 error)More samples if sample variance is larger.Less samples if differen
2021-11-04 10:00:01
327
原创 Hypothesis testing
Example of proportion1 ProportionExperiment: test color color of a buttonClick through probability: N(users who clicked) / N(total users)1000 users in both control and treatment groupsResults:Control group: 1.1% CTPTreatment group: 2.3% CTPSign
2021-11-04 04:17:35
577
原创 reduce/map/combination/permutation - itertools Python example 常用python itertools 函数
Table of ContentsMap/ Reducecombination/ permutationMap/ Reducefrom functools import reducecombination/ permutationIf we want to input parameterlist(map(lambda s: list(combinations(s, 2)), edge_lst))
2021-11-02 12:27:04
121
原创 Review 4995 Applied Machine Learning
Table of contentsLecture 1Basic conceptExploratory Data Analysis & VisualizationLecture 2Supervised Learningk-nearest neighborsDevelopment-test splitRandom SplittingStratified SplittingStructured SplittingHyperparameter tuningModel selectionModel selec
2021-10-27 12:17:15
966
原创 Machine Learning Review Note
目录Bagging, Boosting, StackingTF-IDFBagging, Boosting, Stackinghttps://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205https://zhuanlan.zhihu.com/p/36822575The main hypothesis is that when weak models are correctl
2021-09-23 01:30:43
1768
原创 【量化笔记】波动volatility相关技术指标以其含义
目录Average True Range (ATR)计算操作Bollinger Bands计算操作Bollinger BandWidth%BUlcer Index计算Average True Range (ATR)参考:https://www.thebalance.com/how-average-true-range-atr-can-improve-trading-4154923https://school.stockcharts.com/doku.php?id=technical_indicator
2021-06-10 18:08:47
2835
1
原创 【量化笔记】动量Momentum相关技术指标以其含义
目录Awesome Oscillator (AO)计算Kaufman's Adaptive Moving Average (KAMA)计算Step 1: Efficiency Ratio (ER)Step 2: Smoothing Constant (SC)Step 3: KAMAPercentage Price Oscillator (PPO)计算Percentage Volume Oscillator (PVO)Rate-of-Change (ROC)Relative Strength Index (R
2021-06-05 16:43:57
2867
原创 【量化笔记】移动均线
目录什么是移动平均线了解移动均线两种移动均线计算公式什么是移动平均线在统计中,移动平均值是一种计算,用于通过创建整个数据集的不同子集的一系列平均值来分析数据点。在金融中,移动平均线(MA)是技术分析中常用的股票指标。计算股票移动平均线的原因是通过创建不断更新的平均价格来帮助平滑价格数据。通过计算移动平均数,可以减轻指定时间范围内随机短期波动对股票价格的影响。指数移动平均线(EMA)是加权平均,它对最近几天的股票价格给予更大的重视,使其成为对新信息更敏感的指标。了解移动均线移动平均线是一种简单的
2021-05-21 16:50:26
1036
原创 【量化笔记】动量
目录动量(Momentum)定义了解动量动量交易的缺点动量(Momentum)定义动量是证券价格加速的速度,即价格变化的速度。动量交易是一种策略,趋势回升时进入趋势。简而言之,动量是指价格趋势在特定时间内持续上升或下降的惯性,通常同时考虑价格和数量信息。在技术分析中,动量通常通过振荡器(oscillator)进行测量,并用于帮助确定趋势。了解动量势头强劲的投资者喜欢追逐业绩。他们试图通过投资以一种或另一种趋势发展的股票来获得alpha回报。上升趋势的股票称为热门股票。有些比其他的更热(通过一段时
2021-05-18 20:02:03
1109
原创 【量化笔记】技术指标 Technical Indicator
目录什么是技术指标技术指标如何工作技术指标类型什么是技术指标根据价格、交易量、未平仓合约等数据,构造的启发式(heuristic)、或者基于模式(pattern-based)的信号。交易员可以通过这些信号来预测未来的价格走势。很多技术指标可以大致上分成两类:overlays(叠加)和oscillators(震荡)技术指标如何工作技术分析是一种交易学科,用于通过分析从交易活动中收集到的统计趋势(例如价格变动和交易量)来评估投资并确定交易机会。与基础分析师试图根据财务或经济数据评估证券的内在价值不同
2021-05-18 16:07:33
861
1
原创 Macbook Pro/MacOS安装LightGBM
直接使用conda就能够安装conda install lightgbmLightGBM官网上写着可以直接使用brew安装,如下brew install lightgbm然而安装成功后,却不能在python环境中正常import,希望有明白人能解答一下参考资料:https://towardsdatascience.com/install-xgboost-and-lightgbm-on-apple-m1-macs-cb75180a2ddahttps://lightgbm.readthedoc
2021-05-15 15:11:31
226
Video_Games.csv
2021-05-27
https://raw.githubusercontent.com/selva86/datasets/master/a10.csv
2020-12-29
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人