人工智能工具包标志二。
aitools的Python项目详细描述
工具
人工智能工具包标记2。
- 免费软件:apache软件许可证2.0
- 文档:https://aitools.readthedocs.io。
功能
n克
- Predict the next word of the sentence that further can be use for predicting the entire sentence.
- Generation of text like a paragraph/article on Health or environment or any other topic (given relevant data).
- Generation of the next GOT books!!!.
- Next Phase - Use of bayesian statistics and more nice probability models for getting the best results.
查找:demo文件n_gram.py
k-平均值
- Creates clusters and classify new nodes into those clusters.
- Efficient and easy to implement.
- Next Phase - Multiple algorithms implementation.
查找:demo file k_means.py和k_means_color_cluster.py
逻辑回归
- From probability of event happening or not to the next Data point, it is a very versatile algorithm.
- Simple Implementation and Training routines.
- Uses Gradient Descent and Sigmoid Function.
- Next Phase - Improvements in Gradient Descent is still in progress.
查找:demo file logistic_regression.py
朴素的贝叶斯
- Algorithm based on Bayesian statistics, which is capable of creating classifications like an intent classifier or sentiment analysis engine(A true complex one not the Pathetic Twitter thing).
- Created using little bit of probability and bayesian statistics with a hint of programming and with a lot of love.
- Next Phase - Flexibility and more control on different algorithm.
查找:演示文件naive_bayes.py
决策树
- Algorithm based on Gini logic and info gain for prediction.
- Trees can be very use full but may lead to aver fitting easily Be Cautious.
- Next Phase - Flexibility and more control on different algorithm.
查找:demo file decision_tree.py
预处理
- Utils for Pre Processing Data.
- Next Phase - More Cool Functions.
查找:demo file pre_processing.py
其他
- In build Mathematics Util.
- Vector Math, Probability, and some Statistic
- Util will be upgraded as per the need.
- Gradient Descent.
- Entire Gradient Descent algorithm combined into one.
- Usability Docs will be Provided Soon.
下一次更新
- Improving all algorithms and giving proper documentation on usage.
- Providing more user control on algorithm selection and output handling.
- Detailed Demo files.
- Sudo codes and algorithm explanations.
- Will be available to lower version( > 2.7).
学分
这个包是用Cookiecutter和audreyr/cookiecutter-pypackage项目模板创建的。
历史记录
0.1.0(2019-04-14)
- GitHub安装程序。
- 添加了Travis CI Build。
- 在readthedocs中添加了文档框架。
- pypi上的第一个版本。
0.1.0(2019-04-21)
- N-gram算法版本1已完成。
- 一级文档。
- 添加了N-gram演示文件。
0.1.0(2019-04-25)
- 创建渐变下降。
- 添加了词干分析器(仍然需要改进,并且在整个项目中没有使用)。
0.1.0(2019-04-28)
- 逻辑回归补充。
- 添加了演示文件。
0.1.2(2019-05-01)
- 天真的贝耶斯创造了。
0.1.2(2019-05-01)
- 添加了预处理。
0.1.2(2019-05-01)
- 创建的第一级文档。
0.1.2(2019-05-27)
- 失去了大部分东西的踪迹。
- 最后我们有以下内容
- K表示
- 逻辑回归
- n克
- 天真的贝耶斯
- 预处理
- SVM正在进行中。
0.1.3(2019-06-24)
- 添加决策树(基尼法和杂质法)。
- 先前的算法改进正在进行中。