Machine Learning is Fun – Part 16 – Recommendation System Part 2

Viewer Rating
[Total: 2    Average: 4.5/5]
In my last post, I explained the concept and design of recommendation system for movie based products.

Today, I want to explain the model for hotel based recommendation system.

  1. You must collect demographic data of customers who have booked the hotels and you can get it easily from any online hotel-booking website like makemytrip or booking.com etc.
  2. Collect the details of hotels like room, ambience, hotel star marks, longitude, latitude, hotel location, gym, swimming pool, terrace, local cab book facility etc etc.
  3. We need the customer feedback like review, rating etc.

Data Preparation

  1. After collecting the 3 sets of data, we shall load it in a database or Python data frame.
  2. We shall remove the rows if any null value is found in any column.
  3. Check the categorical variables and convert to numeric using one hot encoding.
  4. Convert the customer feedback text values into some numeric value using sentiment analysis (a separate topic).
  5. Now, we can go for content based recommendation model using rating and feedback point.
  6. As we have longitude and latitude for hotels, we can use it in k-means clustering to group the hotels which have feedback or rating > avg value of rating.
  7. We should use the model as location based recommendation.

Hope you like my post and stay tuned for the next interesting discussion.

1 thought on “Machine Learning is Fun – Part 16 – Recommendation System Part 2”

Leave a Comment

%d bloggers like this: