For generating recommendation, at first, we check the user profile to find his / her preferences in non-spatial features such as price and rating. We, then, remove the hotels those do not satisfy user’s preferences over the non-spatial features.
Then, we compute similarity between the surrounding environments of the remaining hotels with the user’s preferences on the surrounding environments using cosine similarity measure. We can then recommend top-k hotels to the user in the decreasing order of their scores.
Now, consider user u1’s preferences in the surrounding facilities and 10 different hotels surrounding environments as shown in Table IV. Using the cosine similarity measure, we found the similarity scores between user u1 and 10 different hotels as shown in last column of Table IV. Considering the similarity scores of Table IV, our system recommends hotels 797 and 102 as top-2 hotels.