Sample evaluation | ideaforge

Example · federated fraud detection

There's a real idea here. It isn't proven yet.

Every evaluation shows what is strong, what would raise the score and what to do next.

Asset Score

46.5

Out of 100, from twelve pillars · a starting point for development

Verdict: Not recommended

Feasibility (E10) is 22, under the hard hurdle of 40. AI reads the inputs; the score itself is a fixed calculation.

What is strong

The problem is specific enough to evaluate.

What would raise the score

A demonstration of the mechanism would change this example.

What to do next

Show that the mechanism works before anything else.

Problem and ValueE1, E6 · 18%

65.6/100The problem is specific; who benefits is named, not measured.

Solution and FeasibilityE2, E10 · 20%

36.0/100The mechanism is described, not demonstrated. Needs proof

Differentiation and DefensibilityE4, E5 · 20%

48.5/100Nearby work exists; the difference is not proven.

Opportunity and TimingE3, E7, E11 · 19%

49.2/100A path to market is sketched, not evidenced.

Evidence and Asset MaturityE8, E9, E12 · 23%

36.5/100The record is too thin to treat as ready. Needs proof

A similar detection approach

Example A · web search

Moderate overlap · conceptual

A shared-model fraud product

Example B · product search

High overlap · market

A federated analytics paper

Example C · web search

Moderate overlap · implementation

Where we looked

Patents2 kept as close

Research papersSearch failed · a visible gap

Products1 kept as close

Web2 kept as close

A failed search is never shown as “nothing similar exists”.

What we read

Problem, solution and audienceExample text

FilesNone in this example