Mogu: a Collection of Simple Numerical Routines
This is a numerical packages collecting various routines of numerical algorithms. This is not only for convenience, but also demonstration.
I tried to make each file to be independent of each other. If there are functions that are universally used, they will be put into the package
mogu.util. However, I try my best to keep the number of these functions as few as possible.
You must have
numpy pre-installed. GCC and GFORTRAN compilers have to be available in the machine.
After ensuring they have been installed, type the following to install
>>> pip install -U mogu
- Association rule using apriori algorithm;
- Binomial tree algorithm for European and American options pricing;
- Exponential and sigmoid curve fitting;
- Simulated voltage for networks; (moved to new package graphflow since release 0.1.12)
- Google Page rank; (moved to new package graphflow since release 0.1.12)
- Voter rank: Wilson's score;
- Dynamic programming: Damerau-Levenshtein distance;
- Topological data analysis; (implementation moved to
moguTDAsince release 0.1.13)
- Gini coefficients;
- Multivariate Gaussian distribution sampling;
- Probability crosswalk;
- Tensor decomposition using Jennrich algorithm and alternating lease square (ALS) method;
- Discrete Fourier transform (DFT);
- Fast Fourier transform (FFT).