Commodity Documents — Supply Demand Fields

Bilby assesses whether the information contained in a document is likely to affect the supply, demand, or neither for any relevant commodity. If such an impact is detected, we also quantify its likelihood.

General notes: The fields demand_supply_impact_probability_impact, demand_supply_prediction, demand_probability and supply_probability are related in the following three ways. Namely, if the value of demand_supply_impact_probability_impact is no impact, then:

1. The value of demand_supply_prediction is necessarily no prediction, and;
2. Both demand_probability and supply_probability have value NaN.

However, if the value of demand_supply_impact_probability_impact is impact, then:

3. The demand_probability and supply_probability fields form a discrete probability distribution, hence sum to 1.0.

---

demand_supply_impact_prediction

Definition: A binary prediction model that decides if the information contained in the underlying document will impact either:

1. One of the supply or demand of the relevant commodity (in which case the prediction is impact), or;
2. Neither the supply nor demand of the relevant commodity (in which case the prediction is no impact).

Possible values: One of the two class labels: impact or no impact.

---

demand_supply_impact_probability_impact

Definition: The probability that the information contained in the underlying document will impact EITHER the supply OR the demand of the relevant commodity.

Possible values: A float in the interval [0, 1] (i.e., between 0 and 1, inclusive).

Note: The pair of fields:

(demand_supply_impact_probability_impact, demand_supply_impact_probability_no_impact)

form a discrete probability distribution, hence sum to 1.0.

---

demand_supply_impact_probability_no_impact

Definition: The probability that the information contained in the underlying document will impact NEITHER the supply NOR the demand of the relevant commodity.

Possible values: A float in the interval [0, 1] (i.e., between 0 and 1, inclusive).

Note: The pair of fields:

(demand_supply_impact_probability_impact, demand_supply_impact_probability_no_impact)

form a discrete probability distribution, hence sum to 1.0.

---

demand_supply_prediction

Definition: Whether the underlying document affects the demand, supply or neither, of the relevant commodity.

Possible values: One of the three class labels: demand, supply or no prediction.

---

demand_probability

Definition: The probability that the underlying document affects the demand of the relevant commodity.

Possible values: There are two cases:

1. If demand_supply_impact_prediction has value impact, then demand_probability is a float in the interval [0, 1] (i.e., between 0 and 1, inclusive).
2. If demand_supply_impact_prediction has value no impact, then demand_probability has value NaN.

Note: As mentioned in the general notes at the top of this page, if the value of demand_supply_impact_probability_impact is impact, then the fields demand_probability and supply_probability form a discrete probability distribution, hence sum to 1.0.

---

supply_probability

Definition: The probability that the underlying document affects the supply of the relevant commodity.

Possible values: There are two cases:

1. If demand_supply_impact_prediction has value impact, then supply_probability is a float in the interval [0, 1] (i.e., between 0 and 1, inclusive).
2. If demand_supply_impact_prediction has value no impact, then demand_probability has value NaN.

Note: As mentioned in the general notes at the top of this page, if the value of demand_supply_impact_probability_impact is impact, then the fields demand_probability and supply_probability form a discrete probability distribution, hence sum to 1.0.