This repo is a fork from FKD13's 12urenloop-hmmlearn repo, which was used for the course Machine Learning. It's now used for developing lapper algorithms.
If you want to write your own algorithm, start with additional_new_algorithm_testbench.ipynb. Use the data in data_reimport.zip (and move the folder to data).
A good simple lapper was implemented in additional_new_algorithm_testbench.ipynb, it's faster, simpler and less error prone than Viterbi.
Content from FKD13's repo below:
The java part of this project can be found in the HMM.java file. Note this code won't be runnable but is here for reference.
The Python part of my report is split into 3 notebooks matching the corresponding section in the report:
- 3.1.fitting_the_data.ipynb
- 3.2.improving_the_data.ipynb
- 1.3.viterbi_lapper.ipynb
The required data to run the notebooks is provided in the data and data_reimport folders.
If only the archives are available than these should be unpacked.