Discover @stxupengyu Open Source projects

Pangyu (@stxupengyu)

stxupengyu

Stars
594
Global Rank 49,761 (Top 2 %)
Followers 126
Following 47
Registered about 6 years ago
Most used languages

Jupyter Notebook
74.4 %

MATLAB
11.6 %

Python
9.3 %

R
4.7 %
Location 🇨🇳 China
Country Total Rank 10,526
Country Ranking

MATLAB
42

Jupyter Notebook
283

R
461

Python
9,683

BP-RBF-Prediction

使用BP神经网络、RBF神经网络以及PSO优化的RBF神经网络进行数据的预测

PSO-RBF-NN

使用粒子群算法优化的RBF神经网络进行预测。RBF neural network optimized by particle swarm optimization is used for prediction.

Air-Quality-Prediction

2021年研究生数学建模竞赛B题，全国二等奖，空气质量预报二次建模，时间序列数据分析与回归预测。Time Series Prediction&Air Quality Prediction.

Jupyter Notebook

multi-factor-strategy-joinquant

在聚宽（joinquant）平台上使用多因子策略进行量化投资模拟。

Jupyter Notebook

RBF-BP-MATLAB

使用RBF、BP神经网络进行预测。RBF/BP neural network is used for prediction.

Stochastic-Process-Ross-2nd-edition

Here is the exercise solution of stochastic process Ross 2nd Edition collected by the author. The answers are from the stochastic process courses of Umich, Columbia University and BJTU respectively. Due to the different assignments assigned by each teacher, the answers provided by each university are not complete, for your comprehensive reference.

BPNN-MATLAB

使用bp神经网络预测电力负荷，使用小型数据集，通过一个简单的例子。Using BPNN to predict power load, using small data set, a simple example.

LSTM-regression-and-classification

使用LSTM对股票价格进行回归预测，对股价涨跌进行分类预测。We use LSTM to forecast the stock price and classify the rise and fall of the stock price.

Jupyter Notebook

time-series-analysis

使用经典的AR、MA、ARMA、ARIMA、ARCH、GARCH时间序列模型进行模型的检验和拟合。The classic AR, MA, ARMA, ARIMA, ARCH, GARCH time series models are used to test and predict the model.

Jupyter Notebook

Grey-Model

使用灰色系统理论做负荷预测。Using Grey System Theory to Make Load Forecasting

Chinese-Sentiment-Analysis-and-LDA-Topic

使用中文情感词汇本体库进行情感分析，之后对每种情感的文本进行主题分析。Using Chinese Sentiment Dictionary for Sensitive Analysis, Then applying LDA Topic Analysis for each Emotion.

Jupyter Notebook

Econometrics-Example

计量经济学的实例分析包括多元回归分析，多重共线性，对数回归，虚拟变量分段线性回归，多项式拟合以及时间序列。The case analysis of econometrics includes multiple regression analysis, multicollinearity, logarithm regression, piecewise linear regression of dummy variable, polynomial fitting and time series.

Jupyter Notebook

LSFA

The code for our paper "Label-Specific Feature Augmentation for Long-Tailed Multi-Label Text Classification”

Random-Forest-Parameter-Selection

通过十折交叉验证进行参数选择，最后利用最优参数进行随机森林回归预测。Through ten fold cross validation, the parameters were selected, and finally the optimal parameters were used for random forest regression prediction.

Jupyter Notebook

imbalanced-classification

根据60个特征，70万条数据预测5G用户，一个典型的不平衡二分类问题。According to 60 features, 700000 pieces of data predict 5G users, a typical imbalance problem.

Jupyter Notebook

Multi-LSTM-for-Regression

使用LSTM处理回归问题，每个输入特征的时间窗维度不一样，因此，也可以看作利用了多个LSTM特征提取器。When LSTM is used to deal with regression problems, the time window dimension of each input feature is different. Therefore, it can also be regarded as using multiple LSTM feature extractors.

Jupyter Notebook

GMM-KMeans-for-Outlier-Detection

针对一维时间序列数据，采用GMM和K-Means算法进行异常点检测。For one-dimensional time series data, GMM and K-means algorithm are used to detect outliers.

Jupyter Notebook

LDA-for-Chinese-Topic-Generation

使用LDA模型进行中文文本的主题生成。Using LDA Model for Chinese Topic Generation.

Jupyter Notebook

SVM-RFE

SVM classification, RFE feature selection

CVPR-2020-LEAP

Unofficial implement of LEAP(Deep Representation Learning on Long-tailed Data: A Learnable Embedding Augmentation Perspective) for Multi-Label Classification.

P300-BCI-Data-Analysis

2020年研究生数学建模竞赛C题，全国二等奖，分析脑机接口数据进行分析预测。The data of BCI were analyzed and predicted.

Jupyter Notebook

ARIMA-Plot-of-Residuals

使用AIC准则进行参数选择，之后采用ARIMA模型进行时间序列预测，最后给出残差图。The AIC criterion is used to select the parameters, and then ARIMA model is used to predict the time series. Finally, the residual diagram is given.

Jupyter Notebook

NCF-MF-for-Recommendation

分别使用传统方法（KNN，SVD，NMF等）和深度方法（NCF）进行推荐系统的评分预测。Traditional methods (KNN, SVD, NMF, etc.) and depth method (NCF) were used to predict rating of the recommendation system.

Jupyter Notebook

Credit-Data-Analysis

实现对信贷数据的数据预处理，数据分析。之后利用多种分类算法对公司是否违约进行预测。Realize the data preprocessing and data analysis of credit data. Then, it uses a variety of classification algorithms to predict whether the company defaults.

Jupyter Notebook

SARIMA

使用SARIMA模型进行时间序列预测。Time series prediction using SARIMA model.

Jupyter Notebook

sklearn-regression-algorithm

常见sklearn回归算法（随机森林，adaboost，bagging，knn等）在示例数据集上的使用。The application of common sklearn regression algorithms (random forest, AdaBoost, bagging, KNN, etc.) on the sample dataset.

Jupyter Notebook

MF-for-Movie-Recommendation

使用矩阵分解方法进行电影推荐的评分预测。The matrix factorization method is used to predict the rating of movie recommendation.

Jupyter Notebook

Yelp-Recomendation-Algorithms

在Yelp数据集上摘取部分评分数据进行多种推荐算法（SVD,SVDPP,PMF,NMF）的性能对比。Some rating data are extracted from yelp dataset to compare the performance of various recommendation algorithms(SVD,SVDPP,PMF,NMF).

Jupyter Notebook

maximum-likelihood-estimation

介绍和举例（正态分布、泊松分布、伽马分布）展示了极大似然估计。This paper introduces and gives examples (normal distribution, Poisson distribution, gamma distribution) to show the MLE.

Jupyter Notebook

sparse-convex-clustering-demonstration

R语言包“稀疏凸聚类（scvxclustr）”的实验演示。Experimental demonstration of R package: "sparse-convex-clustering(scvxclustr)".

Jupyter Notebook

TF-IDF-for-Chinese-Keywords-Generation

使用TF-IDF算法进行中文关键词生成任务。Using TF-IDF Algorithm to Generate Chinese Keywords.

Jupyter Notebook

Matrix-Factorization-for-Recommendation

Using Matrix Factorization/Probabilistic Matrix Factorization to solve Recommendation。矩阵分解进行推荐系统算法。

shadowsocksR-Installation-package

NCF-for-Implicit-Feedback

Neural collaborative filtering (NCF) method is used for Microsoft MIND news recommendation dataset.

Jupyter Notebook

lstm-classification

使用LSTM解决分类问题。Using LSTM to solve classification problems.

Jupyter Notebook

Matrix-Factorization-Implicit-Feedback

使用矩阵分解算法处理隐式反馈数据，并进行Top-N推荐。The matrix factorization algorithm is used to process the implicit feedback data and make top-N recommendation.

LDA-gensim

使用LDA模型提取n个句子的主题，并统计每个主题出现的频次。LDA model is used to extract the topic of N sentences, and the frequency of each topic is counted.

Jupyter Notebook

Comment-Sentiment-Analysis

使用基于情感词典的情感分析方法对评论信息进行情感分析。The Sentiment Analysis Method based on Sentiment Dictionary is used for Comment Information.

Jupyter Notebook

Random-Forest-Regression

使用随机森林算法对企业评级进行预测。The random forest algorithm is used to predict the enterprise rating.

Jupyter Notebook

XDA

Official code for our paper "Taming Prompt-based Data Augmentation for Extreme Multi-Label Text Classification”.

Bayesian-Computation-with-R-Solutions

Part of the solutions about 《Bayesian Computation with R》(Jim Albert)

BPNN-R

Using back propagation neural network(BPNN) to forecasting the price. 使用R语言实现BP神经网络回归预测。

Housing-Price-Prediction

Using 5 models(lasso,elastic net,KernelRidge, boosting, xgboost) to predict the housing price

Jupyter Notebook

XML-data-processing

处理多个微博上爬取的XML数据，转换为pandas.dataframe格式。Process XML data crawled from multiple microblogs and convert it to pandas.dataframe format.

Jupyter Notebook

XML-to-DataFrame

将XML通过ElementTree转化为numpy/DataFrame格式。通过一个简单的例子。XML to Numpy/Pandas, a simply example.

Jupyter Notebook

logistic-regression-variable-selection

使用logistic回归预测违约事件，实现变量选择。

Jupyter Notebook

Multi-Label-Classification-Data-Preprocessing

对于多标签分类数据集的预处理。Data preprocessing for multi label classification.

Jupyter Notebook