This document was generated from CDN thread
Created by: Kenneth Leung on 18-08-2010 05:25:21 PM
Just started using the filter and got a tutorial on the training function. I am not sure it is that intuitive.
As I understand it, the filter engine is separate from the data feeds so that you can set up one filter, train it, and then apply it to different feeds. You can use filters in a non training mode, i.e. have the filter and no rating.
As I try to use this, I am checking out the UI.
#1 I am trying to think of a scenario when I would not be in training mode. Since content is dynamically changing in social media, I would always want to rate something up or down. Just like in regular social media sites like facebook, you are always given the option to rate up and down.
#2 If I rate something in the filter to be removed, shouldn't it automatically delete?
#3 When I rate something to be filtered it has the word it as Trained Out with when the feed refreshes, the Trained Out tag disappeared.
Subject: RE: Feedback on the FIlter and training/non training mode
Replied by: Tod Famous on 11-09-2010 10:47:16 PM
1. there may be some environments where there is a formal customer care agent downstream in the workflow perhaps who doesn't understand (or need to understand) how Social Contacts are filtered; therefore the training UI is not relevant important too them the tool allows it to be "hidden" from this type of user by going out of training mode.
For example, we have expect some users to only view the Campaign Results gadget with a specific tag. they would only get work that was traiged by someone else. Perhaps there is a PR/Marketing person that is looking at the full stream and they are tagging work as "support" and there is a support person who only see's work show up when it's tagged support. That support person would not need the complexity of the trianing mode information.
2. Yes, ultimately this makes sense. We implemented training as a distinct operation as a means to first "experiment" with the bayesian filter technology. Once we're comfortable that it's working well, we plan to offer a "auto-train" mode which would train without the up/down arrows. For example, we'd train-in and SC's that marked handled and we would train-out any SC's that are discarded.
3. Yes, the trained out message is a response that indicates success of training. We do not permanently mark the SC as "trained out". We have had other users ask that we perserve this as a state so it is an enhancement we have on our backlog.
