The model developed using TF-IDF showed the best performance, followed by the model using word occurrence and lastly the model using binary frequency.
A possible reason for the better prediction performance of the model developed using TF-IDF is that those words that occur less frequently play a key role in the classification of 'useful' and 'non-useful' articles.
SVM models developed using TF-IDF had better prediction performance compared to those developed using either word occurrence or binary frequency.
The result of our method using both JacS and tf-idf is better than the results of a standard method using tf-idf only.
On the contrary, the worst separation of classes was observed when using tf-idf weighting and Jensen-Shannon divergence.
In our experiment with an image collection with ground truth, we show that the accuracy of JacS alone is sometimes better than the accuracy when using tf-idf alone and that the accuracy of JacS used together with tf-idf is always much better than using JacS or tf-idf alone.
We set (n=4) and use the range of (4,4) to compute normalized n-grams (using TF-IDF) of each advertisement content.
We propose using Jaccard Similarity (JacS), which is also known as Jaccard Similarity Coefficient, for calculating image pair similarity in addition to using tf-idf.
In this paper, we implement keyword extraction methods with different term weights using TF-IDF [7], BM25 [32], and TextRank[30] as our baselines.
For example, if a user is located in Portsmouth, Hampshire, UK, the system will go through similar process as Keywords Criterion using TF-IDF.
