Influence Score

Advantages

Compared to the WoT Score, the Influence Score possesses the following advantages: - Unlike the WoT Score, the Influence Score can see more than 2 hops away on your social graph. - Unlike the WoT Score, the Influence Score avoids being a popularity contest through the use of this method. - Through the method of proxy trust data interpretation, the Influence Score is well suited for the synthesis of trust data from multiple sources into a single score. - The Influence Score handles contextual trust naturally through the calculation of Contextual Influence Scores.

Disadvantages

The primary disadvantage to the calculation of Influence Scores is the high computational requirement. Unlike the WoT Score, the Influence Score typically cannot be calculated “on the fly” but must be calculated ahead of time.

A common objection to the Influence Score is that it is “too complicated.” In practice, this is not the case, neither for developers nor for users.

It may be argued that an additional impediment to adoption of the Influence Score method is the fact that many developers have high WoT Scores and may therefore be disincentivized to explore alternative scoring systems. It remains to be seen whether there is any merit to this argument.

Method of Calculation

The concept that forms the foundation of the Influence Score is the concept of weights. Average scores, for example, are calculated as weighted averages rather than simple averages. Once the decision to use weights is made, the details of the method fall inevitably into place.

Weights are functions of one or more variables. Typically, the weight is proportional to the relevant Influence Score of the author of the piece of data in question (follow, like, zap, etc). Often, the weight is also proportional to confidence, which may be inferred through the interpretation of proxy data or stated explicitly by the author.

Influence Scores can be contextual. Contexts can be managed in a decentralized manner, may be fine grained or coarse grained, and may be organized into hierarchies. Typically, Influence is “inherited” from parent context to child context.