WEEK 6
The new inputs for this research are applied every two weeks. My first input was “how to make a pizza”. After organizing the data that I collected in the last 2 weeks, the similar pattern to the previous data pattern showed. Before putting the input in, there are almost no related videos in the algorithm. After putting the input into the search bar, there is 95% or more similarity to the input in the algorithm. So, I assumed this experiment may end up being very similar to the results from the previous research. However, I found a surprising part while gathering data from the second input, “how a hummingbird flies”. There was 0% similarity to the input before searching as similar to the first input. But there was 80% similarity to the input after searching, which is surprising to me since the results after searching tend to show 90% or above in similarity in the previous research and the first input. There were four unrelated videos to inputs after searching. The four videos are as followed:
1: A random video
2: Different bird flying
3: Hummingbird nectar recipe
4: Tips for purchasing
Currently, I am not sure why the similarity
has decreased from the first input to the second input. My guess on this is
that the first unrelated video would be a random video that YouTube put it in
to find what I really want to do since I don’t watch any of the videos that I
search on this device. And the other related videos are not matched to the input
completely, so I assume that changing inputs specifically would give some
changes in similarity. I am looking forward to seeing more changes in the
algorithm.
Comments
Post a Comment