The trend analysis component, upon receiving this message, would analyze it and store it in a knowledge base. You may assume to have a poster table containing poster ID, poster type (keywords such as multimedia, electronic games, etc.). Therefore, the trend analysis component can store in the knowledge base the following information: poster type, count, year.
When the voting is done, a message is sent to the voting component to close the vote. The voting component will send a message to the trends analysis component to complete current year's statistics and perform the prediction.
Trends analysis component can be used to predict what percentage of voters would vote for what type of posters, based upon previous years' trends. The simplest prediction is to take last year's entry from the knowledge base. A more sophisticated prediction may take the most recent two/three/four year's average, and so on.
So, you need to specify several type of new messages, and come up with an example (similar to the one outlined above) to describe the scenario.