This Repository contains the source code of Gezirapal project a deep learning project that helps people grow crops by recommending the type of crops to cultivate, check the contribution for more information contact: gezirapal@gmail.com
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
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.
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