The first multi-modal machine learning model which rates customer service friendliness 💑
To install and use the project, you must have the latest version of Python, Anaconda, and Jupyter Notebook installed on your device. Links to both are provided below:
-
- Although having Anaconda is not necessary, it is reccommended to download in order to run our program on a seperate virtual environment. This is to allow easier access to download and delete python libraries without the hassle of directly downloading them to the base environment (aka the local device) and potentially take up space on the memory. To do this, open the Anaconda command line interface and type
conda activate rizzerator. From here, you would runjupyter notebookand locate the directory containing the files. - If Anaconda is not downloaded, run the program on any other IDE of your liking (we reccomend VS Code) :)
- To install different Python libraries, you can run
pip installfollowed by the library you choose to download or, if you downloaded Anaconda, you can runconda installfollowed by the library of your choosing. - Note: some Python libraries might not have
conda install, so its still fine to run `pip install' while using Anaconda.
- Although having Anaconda is not necessary, it is reccommended to download in order to run our program on a seperate virtual environment. This is to allow easier access to download and delete python libraries without the hassle of directly downloading them to the base environment (aka the local device) and potentially take up space on the memory. To do this, open the Anaconda command line interface and type
After downloading all the necessary tech stack above, you can successfully run the FinalRizz.ipynb file. To do this run each individual block of code in order from top to bottom to finalize before you reach the classification models. Run the classification models and review the results that are produced. Data on accuracy, precision, f1 score, and recall will be shown for each model as well as their respective confusion matrices.
Others can contribute to the project by:
- Feeding the model more training data in order for it to better predict the correct results.
- Including aditional modalities that seem fit for detecting customer service friendliness.
- Keeping track of any new issues added to the GitHub repository.
If you have any questions, concerns, or contributions contact one of the authors:
❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️Thank You in Advance!!!❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️❤️