In order to show the belief propagation of our new ranking algorithm, along the manifold of the data, we generated a set of toy data, selected one data point as a query sample, and queried the rest samples to rank them.
At dozens of stalls, shoppers reached for raw garlic samples and queried farmers about New York's growing conditions.
Subsequently, a new classifier is designed which utilizes CGM to find k nearest neighbors from each class for the query sample and then classifies this query sample to the class whose cognitive gravitation is largest.
For data classification applications, the RBF network has one output node corresponding to one class of samples and a query sample is predicted to belong to the class of which the corresponding output node yields the maximum value.
We applied this method in two phases, the first a verification phase in which we sought to rediscover known exposures applied to samples, and a query phase, in which we sought to find factors associated with cancer gene expression datasets.
The basic idea is to find close neighbors to a query sample and train a local support vector machine that preserves the distance function on the collection of neighbors.
These spectrogram images, along with their related song information would be stored for searching and retrieval of the query sample and its corresponding match.
For the experimental prototype, a Brute-Force matcher is used to match the corresponding descriptors between the music query sample and the database of spectrogram images.
By considering this relevance feedback, the user can continue with the next step by defining the next part of the query sample and can thereby identify a highly frequent pattern, which was initially not known.
When the descriptors are matched, a threshold was used for the distances of the matches, if the distance is far apart between descriptor matches of the query sample and a database song then it may have been a false positive detected.
