Skip to main content
Noonum turns unstructured text (filings, news, transcripts, and research) into investable signals. A company’s exposure to an investment theme shows up in the language it and the world use to describe it. Noonum reads that language across a universe of companies and produces scores you can rank, weight, and compare. This page covers what each signal represents and how to read it, not how it is produced. Use it to interpret a company’s conviction score, judge how strongly a company ties to its theme, and decide how to weight holdings.

​
Strategies, holdings, and portfolios

Three terms are easy to confuse. Keep them straight:
  • A strategy is a thematic definition. It resolves to holdings: a ranked list of companies, each scored by thematic strength. Most of the API is about strategies and their holdings.
  • A weighted portfolio is the result of running POST /strategies/{strategyId}/optimize on a strategy. Optimization tilts the holdings’ weights by a signal score. It is a computed result you back-test or act on, not a stored object.
  • A saved portfolio is the /portfolios resource: a named, weighted basket of holdings that stands on its own, independent of any strategy. You create one yourself, or save a weighting you computed, then reuse or back-test it. See the API reference for the endpoints.
A common flow ties them together: build a strategy, review its holdings, optimize to a weighted portfolio, back-test it, and optionally save that weighting as a portfolio to reuse later.

​
Concepts

Read these in order, starting with Signals and scores.
  • Signals and scores: what conviction, linguistic beta, and market buzz represent, their value ranges, and how to read them when ranking a strategy.
  • Portfolio optimization: how Noonum turns a strategy’s ranked holdings into a weighted portfolio with score-tilt weighting, tilting weight toward higher-scoring companies.
  • Research papers: the published research behind the methodology, available as PDFs.