| CPC G06F 16/9538 (2019.01) [G06F 16/958 (2019.01); G06Q 30/0643 (2013.01); G06Q 50/163 (2013.01)] | 20 Claims |
|
1. A system for rearranging tags associated with a real estate listing photocard displayed on a graphical user interface (GUI) of a computer system using known and unknown levels of webpage traffic, the system comprising:
at least one processor;
a remote data store storing:
a first set of real estate listing tags, wherein each tag of the first set of real estate listing tags indicates a home attribute and is associated with (i) a geographic region and (ii) an known level of webpage traffic;
a second set of real estate listing tags, wherein each tag of the second set of real estate listing tags indicates a home attribute and is associated with (i) the geographic region and (ii) an unknown level of webpage traffic; and
at least one memory coupled to the at least one processor and storing instructions that, when executed by the at least one processor, perform operations comprising:
accessing the remote data store to obtain (i) the first set of real estate listing tags and (ii) the second set of real estate listing tags;
comparing, for each of the first set of real estate listing tags, the known level of webpage traffic to a threshold level of webpage traffic, to determine whether the respective known level of webpage traffic meets or exceeds the threshold level of webpage traffic;
generating a set of popular real estate listing tags, from the first set of real estate listing tags, by selecting known real estate listing tags from the first set of known real estate listing tags having a respective known level of webpage traffic meeting or exceeding the threshold level of webpage traffic;
performing clustering, by a K-Nearest-Neighbor algorithm, using the set of popular real estate listing tags and the second set of real estate listing tags, to determine a set of clusters, wherein each cluster of the set of clusters indicates semantically similar real estate listing tags, and wherein each cluster of the set of clusters includes at least one of the popular real estate listing tags of the set of popular real estate listing tags and one or more real estate listing tags of the second set of real estate listing tags;
for each cluster of the set of clusters:
generating a cosine similarity distance value between (i) each of the real estate listing tags of the second set of real estate listing tags and (ii) a popular real estate listing tag included in the respective cluster;
determining, for each of real estate listing tag of the second set of real estate listing tags, an estimated level of webpage traffic based on (i) the cosine similarity distance value and (ii) the level of known webpage traffic of the popular real estate listing tag included in the respective cluster;
generating a final set of real estate tags by merging the second set of real estate listing tags with the set of popular real estate listing tags;
ranking each real estate listing tag of the final set of real estate listing tags, in descending order, based on the respective level of known or estimated webpage traffic associated with the respective real estate listing tag;
selecting, from the final set of ranked real estate listing tags, a real estate listing tag having the highest level of known or estimated webpage traffic as compared to all other ranked real estate listing tags of the final set of ranked real estate listing tags; and
generating, for display, on a real estate listing photocard, the selected real estate listing tag based on the determined level of known or estimated webpage traffic, wherein the real estate listing photocard is associated with a real estate listing that corresponds to the geographic region.
|