This year, Twitter had a series of criticisms by distancing himself from that timeline that shows tweets chronologically and began to prioritize them with an algorithm that is now in charge of showing those that Twitter thinks are most interesting to the user.
From today, Twitter will be sorting search results in the same way, hoping that the most relevant results improve the end user experience. A year of changes for that huge number of tweets that are revealed to the user.
It is Lisa Huang, one of the software engineers on the Twitter search team, who explains the reason for the change. The reason for this modification is because Twitter moves fast and the most recent results would not show exactly the tweet you would be looking for.
The team has been testing a number of factors that refine the results as with the machine learning model that currently positions the results. But it's really the keystrokes you make in a tweet what they define for Twitter what is relevant to you.
El behavior of a user Using your Twitter is of great value to try to understand the interaction with the various tweets. It is with this information that Twitter can train the machine learning models that predict how a tweet will be interacted with. Those models then rank relevant tweets based on the likelihood that users will interact with them.
While this makes its own sense, the fact is that those tweets that receive the most retweets they don't necessarily make it relevant. To show the changes, Twitter made a comparison between the old model (the one on the left) and the new one (placed on the right). With the new model, brands and official accounts are favored, while individual users are left lower ...
A novelty that will bring your criticism and that joins another from not so many days ago.