https://github.com/daya-jin/ml_for_learner
Implementations of the machine learning algorithm with Python and numpy
https://github.com/daya-jin/ml_for_learner
implementation machine-learning python3 sklearn
Last synced: 6 months ago
JSON representation
Implementations of the machine learning algorithm with Python and numpy
- Host: GitHub
- URL: https://github.com/daya-jin/ml_for_learner
- Owner: Daya-Jin
- License: apache-2.0
- Created: 2018-12-20T14:02:33.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2021-10-20T10:05:27.000Z (over 4 years ago)
- Last Synced: 2025-01-31T02:21:50.103Z (about 1 year ago)
- Topics: implementation, machine-learning, python3, sklearn
- Language: Jupyter Notebook
- Homepage:
- Size: 7.92 MB
- Stars: 87
- Watchers: 1
- Forks: 44
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# ML_for_learner
该项目旨在使用numpy实现一个类scikit-learn的mini机器学习库，对于相关的知识，均配有blog文章对其理论进行讲解，对于部分功能，还配有notebook分析代码实现上的细节。该项目的初衷是为那些算法学习者提供从理论到实现的一站式服务。
由于本人学识有限，并且没有Python开发经验，该库目前还是一个非常松散的代码集合体。如果你在blog、notebook或者code中发现任何纰漏或bug，甚至是觉得哪写的不通顺，都可以联系我，当然也可以直接在项目页面提issue，谢谢。
QQ: 435248055 | WeChat: QQ435248055 | [Blog](https://daya-jin.github.io/)
---
点击算法名称进入相应Blog了解算法理论，notebook指导如何step-by-step的去实现该算法，code为模块化的代码文件。
注：除非特别说明，各模型所接受的数据格式均为```numpy.ndarray```格式，部分也可接受```List```或者嵌套```List```，除此之外的数据格式本人暂不保证。由于目前的Python type hint还不支持numpy，所以在代码中未说明(感谢微信昵称@Stream的提醒)。
## Supervised learning
|Class|Algorithm|Implementation|Code|
|-|-|-|-|
|Generalized Linear Models|[Linear Regression](https://daya-jin.github.io/2018/09/23/LinearRegression/)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/linear_model/LinearRegression.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/linear_model/LinearRegression.py)|
||[Logistic regression](https://daya-jin.github.io/2018/10/02/LogisticRegression/)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/linear_model/LogisticRegression.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/linear_model/LogisticRegression.py)|
|[Nearest Neighbors](https://daya-jin.github.io/2018/12/29/KNearestNeighbor/)|Nearest Neighbors Classification|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/neighbors/KNN.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/neighbors/KNeighborsClassifier.py)|
|[Naive Bayes](https://daya-jin.github.io/2018/10/04/NaiveBayes/)|Gaussian Naive Bayes|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/naive_bayes/Gaussian%20Naive%20Bayes.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/naive_bayes/GaussianNB.py)|
|[Support Vector Machine](https://daya-jin.github.io/2018/10/17/SupportVectorMachine/)|SVC|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/svm/SMO.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/svm/SVC.py)|
|[Decision Trees](https://daya-jin.github.io/2018/08/10/DecisionTree/)|[ID3 Classification](https://daya-jin.github.io/2018/08/10/DecisionTree/#id3)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/tree/ID3_Clf.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/tree/ID3_Clf.py)|
||ID3 Regression|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/tree/ID3_Reg.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/tree/ID3_Reg.py)|
||[CART Classification](https://daya-jin.github.io/2018/08/10/DecisionTree/#classification-and-regression-tree)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/tree/DecisionTreeClassifier.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/tree/DecisionTreeClassifier.py)|
||CART Regression|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/tree/DecisionTreeRegressor.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/tree/DecisionTreeRegressor.py)|
|[Ensemble methods](https://daya-jin.github.io/2018/08/15/EnsembleLearning/)|[Random Forests Classification](https://daya-jin.github.io/2018/08/15/EnsembleLearning/#random-forest)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/ensemble/RandomForestClassifier.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/ensemble/RandomForestClassifier.py)|
||Random Forests Regression|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/ensemble/RandomForestRegressor.ipynb)|[code]()|
||[AdaBoosting Classification](https://daya-jin.github.io/2018/08/15/EnsembleLearning/#boosting)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/ensemble/AdaBoostClassifier.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/ensemble/AdaBoostClassifier.py)|
## Unsupervised learning
|Class|Algorithm|Implementation|Code|
|-|-|-|-|
|Gaussian mixture models|[Gaussian Mixture](https://daya-jin.github.io/2019/03/15/Gaussian_Mixture_Models/)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/mixture/GaussianMixture.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/mixture/GaussianMixture.py)|
|Clustering|[K-means](https://daya-jin.github.io/2018/09/22/KMeans/)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/cluster/KMeans.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/cluster/KMeans.py)|
||[DBSCAN](https://daya-jin.github.io/2018/08/06/DBSCAN/)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/cluster/DBSCAN.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/cluster/DBSCAN.py)|
|[Association Rules](https://daya-jin.github.io/2018/12/30/AssociationRules/)|Apriori|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/rule/Apriori.ipynb)||
|[Collaborative Filtering](https://daya-jin.github.io/2019/04/03/CollaborativeFiltering/)|[User-based](https://daya-jin.github.io/2019/04/03/CollaborativeFiltering/#user-based-collaborative-filtering)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/recommend/1.%20user_based_CF.ipynb)||
||[Item-based](https://daya-jin.github.io/2019/04/03/CollaborativeFiltering/#item-based-collaborative-filtering)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/recommend/2.%20item_based_CF.ipynb)||
||[LFM](https://daya-jin.github.io/2019/04/03/CollaborativeFiltering/#latent-factor-model)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/recommend/LFM.ipynb)||
## Model selection and evaluation
|Class|Approach|Code|
|-|-|-|
|[Model Selection](https://daya-jin.github.io/2018/12/11/Model_Assessment_and_Selection/)|Dataset Split|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/model_selection/train_test_split.py)|
||[K-Fold](https://github.com/Daya-Jin/ML_for_learner/blob/master/model_selection/KFold.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/model_selection/KFold.py)|
||Stratified K-Fold|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/model_selection/StratifiedKFold.py)|
|[Metrics](https://daya-jin.github.io/2019/03/27/Evaluation_Metircs/)|Accuracy|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/Classification.py#L4)|
||Log loss|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/Classification.py#L53)|
||F1-score|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/Classification.py#L11)|
||[AUC](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/AUC.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/Classification.py#L75)|
||Explained Variance|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/Regression.py#L4)|
||Mean Absolute Error|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/Regression.py#L9)|
||Mean Squared Error|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/Regression.py#L14)|
||R Square|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/Regression.py#L19)|
||[Euclidean Distances](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/pairwise/euclidean_distances.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/metrics/pairwise/euclidean_distances.py)|
## Preprocessing data
|Class|Algorithm|Implementation|Code|
|-|-|-|-|
|[Feature Scaling](https://daya-jin.github.io/2019/03/20/Data_Scaling/)|StandardScaler||[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/preprocessing/StandardScaler.py)|
||MinMaxScaler||[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/preprocessing/MinMaxScaler.py)|
|Unsupervised dimensionality reduction|PCA|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/decomposition/PCA.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/decomposition/PCA.py)|
||SVD|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/decomposition/SVD.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/decomposition/TruncatedSVD.py)|
|Supervised dimensionality reduction|[Linear Discriminant Analysis](https://daya-jin.github.io/2018/12/05/LinearDiscriminantAnalysis/)|[notebook](https://github.com/Daya-Jin/ML_for_learner/blob/master/discriminant_analysis/LinearDiscriminantAnalysis.ipynb)|[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/discriminant_analysis/LinearDiscriminantAnalysis.py)|
|Text Feature|Count Feature||[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/feature_extraction/text.py#L6)|
||[TF-IDF]()||[code](https://github.com/Daya-Jin/ML_for_learner/blob/master/feature_extraction/text.py#L48)|
## Known Issues
整体代码重用性较低。
random forest没有实现并行。
LDA代码存在功能欠缺。
K-Fold代码中使用了```np.append()```，效率较低。