Data source: https://www.kaggle.com/kingburrito666/cannabis-strains
Goal of the model: Take input info from users, including strain type, rating, effects, and flavors, the model will give the recommended strain name.
How does the model work? Step1: Convert the users' input into numbers by using Spacy (Natural Language Processing Package), which is a open-source library for advanced Natural Language Processing (NLP) in Python. Spacy is a package developed by neural network. Step2: NearestNeighbors model is used, which is used to calculate the distance of different vectors. Step3: NearestNeighbors model provides the index of closest strain name, and then the strain name will be extracted from the original dataframe.
The model is saved as a .pkl file, so people who need the model can just load the model by using the pickle.load.