BUGS

[A7med7x7]

in the basic recommender feature, I used ordinary data filtering, as an example. if I select a combination of inputs, the filtering function will return a similar output from the database, which is kind of fast but not efficient.

def predict_possible_crops( district, fertilizer, soil):
    # Filter the dataset based on the provided parameters
    filtered_data = dataset[(dataset['District_string'] == district) &
                            (dataset['Fertilizer'] == fertilizer) &
                            (dataset['soil_string'] == soil)]

the problem when doing doing that type of functionality, the function can return 0 for no combinations found on dataset the solution I'm thinking we can use algorithm that is distance-based, like KNN, to solve this problem by returning the k nearest sample from the feature space

[A7med7x7]

• [ ] we have real issues with dependencies compatibility
• [ ] documentation is not that good ngl