GICS Documents — Information Retrieval Fields
Some documents collected by Bilby are relevant to one or more nodes in the GICS tree (a taxonomy used by analysts to classify equities). Bilby labels such documents with the relevant GICS node, quantifies the degree of this relevance with a floating point score, and assigns ordinal rankings to each day's documents, based on this score.
Field Name | Type | Example/Possible Values | Description |
---|---|---|---|
relevant_gics | string | 2030-Transportation | The GICS node that is relevant to the underlying document. The format is {code}-{name} , where code is the GICS code and name is the GICS name. |
theme_relevance_score | number | 0 ≤ value ≤ 1 | The relevance score of the underlying document to the GICS tree node given by the value relevant_gics . |
theme_relevance_rank | number | 1 , 2 , 3 , or 4 , and so on. | The rank of the underlying document. This is measured by relevance_score , compared to other gics-related documents published on the same day. |
relevant_gics
Definition: The GICS node that is relevant to the underlying document.
Possible values: Any of the nodes on the GICS taxonomy tree. For example:
- (Sector level):
10-Energy
- (Industry group level):
2030-Transportation
- (Industry level):
251020-Automobiles
- (Sub-industry leve;):
30202010-Agricultural products and services
.
theme_relevance_score
Definition: Bilby has created a classifier model for each node on the GICS
tree. Each model computes the relevance of the underlying document to the given
commodity, and assigns this value to the theme_relevance_score
variable.
Possible values: : A float in the interval [0, 1]
(i.e., between 0 and 1,
inclusive).
theme_relevance_rank
Definition: On any given day, the document with the N
th-highest
theme_relevance_score
is assigned a theme_relevance_rank of N
. For example,
the highest-scored document is given a rank of 1; the second-highest, a rank of
2, and so on.
Possible values: Positive integers 1
, 2
, 3
, ...