dsckgec / movierecommendationsystem Goto Github PK

Describe the solution you'd like
Like data_movies dataframe is preprocessed, similarly the features are picked to be picked from movies_data dataframe.

Features to be picked
You need to pick these features from movies_data : 'title', 'tagline', 'genres', 'cast', 'director'.

Describe the solution I'd like
Find out the top 50 most profitable movies, and top 50 least profitable movies and do a horizontal Bar Graph.

Describe the bug
I got a warning that many columns have mixed types and Specify dtype option on import or set low_memory=False.
The solution of setting low_memory = False wont solve the problem .
I have even attached the screenshot , u can see unnamed 3, 4 ,5 6....1262.

May be the problem is with the dataset or i dont know .. please look after this issue .

Expected behaviour
Based on the notebook in this repo , it should have print like this
data_credits columns: Index(['movie_id', 'title', 'cast', 'crew'], dtype='object')

but i got like the below
data_credits columns: Index(['movie_id', 'title', 'cast', 'Unnamed: 3', 'Unnamed: 4', 'Unnamed: 5',
'Unnamed: 6', 'Unnamed: 7', 'Unnamed: 8', 'Unnamed: 9',
...
'Unnamed: 1255', 'Unnamed: 1256', 'Unnamed: 1257', 'Unnamed: 1258',
'Unnamed: 1259', 'Unnamed: 1260', 'Unnamed: 1261', 'Unnamed: 1262',
'Unnamed: 1263', 'Unnamed: 1264'],
dtype='object', length=1265)

Desktop

• OS: [e.g. Windows]
• Browser [e.g. chrome]
• Version [e.g. 10]

Additional context
If this is my mistake please specify the error i have made or please provide solution

Describe the solution you'd like
Apply K-Means Clustering or Gaussian Mixture Model for Recommendation System. You can either use One Hot Encoding or Label Encoding.

Describe the solution I'd like
Apply K-Nearest Neighbor Algorithm to build a revenue prediction system. Use genre_onehot which is One Hot Encoded version of all genre of movies and also only use the movies whose status == released. Also co-relate it with the budget of the movie.

Describe alternatives you've considered
You can also use Linear Regression Model or Logistic Regression model.

The solution I'd like
Apply Cosine Similarity Algorithm on cleaned movies_data

The alternatives you can consider
Find some alternative algorithms than can be helpful for the recommendation system.

Is your feature request related to a problem? Please describe.
There are no heading, no proper alignment, there is extra padding on both sides and the axis are also not named

Describe the solution you'd like
Apply all the things I mentioned and also apply show the value of profit percent beside the respective bars.