Definitions
Key definitions in VIEWS
Prediction outcomes
The VIEWS system currently generates two sets of predictions per type of violence and level of analysis: continuous predictions of the number of fatalities in impending conflict, and
probabilistic dichotomous (conflict/no conflict) forecasts based on the former.
The models produced under the VIEWS1 and VIEWS2 data infrastructures were limited to dichotomous predictions.
Over the course of the Societies at Risk and ANTICIPATE projects, the VIEWS system will also be expanded with predictions for humanitarian impacts of conflict.
The models produced under the VIEWS1 and VIEWS2 data infrastructures were limited to dichotomous predictions.
Over the course of the Societies at Risk and ANTICIPATE projects, the VIEWS system will also be expanded with predictions for humanitarian impacts of conflict.
Continuous predictions (2022 – )
- Predicted number of battle-related deaths (BRDs) per country-month and type of violence, during each of the next 36 months.
- Predicted number of battle-related deaths (BRDs) per PRIO-GRID-month and type of violence, during each of the next 36 months.
Dichotomous predictions (2018 – )
- Predicted probability (0-100%) of at least 25 battle-related deaths per country-month and type of violence, during each of the next 36 months.
- Predicted probability (0-100%) of at least 1 battle-related death per PRIO-GRID-month and type of violence, during each of the next 36 months.
Types of violence
VIEWS applies a “divide and conquer’ strategy to the forecasting problem. It analyses separately 1-3 different outcomes of fatal political violence, as defined
[1] and recorded by the Uppsala Conflict Data Program, (UCDP).
The models developed under the VIEWS1 and VIEWS2 data infrastructures offered predictions for all three types of violence below.
The fatalities model under the VIEWS3 data infrastructure is currently limited to state-based conflict, but is set to expand to non-state and one-sided violence in the near future.
The models developed under the VIEWS1 and VIEWS2 data infrastructures offered predictions for all three types of violence below.
The fatalities model under the VIEWS3 data infrastructure is currently limited to state-based conflict, but is set to expand to non-state and one-sided violence in the near future.
State-based conflict (sb)
Armed conflict between two or more actors – of which at least one is the government of a
state
[2] – over a contested incompatibility
[3] that concerns government and/or territory.
Non-state conflict (ns) (forthcoming)
The use of
armed force between two or more organised armed groups, neither of which is the
government of a
state.
One-sided violence (os) (forthcoming)
The deliberate use of
armed force by the
government of a
state, or by a formally organised group
[4] against civilians.
Notes:
[1] Please note that in order for a given conflict event to be included in the final and annual UCDP-GED dataset that the ViEWS forecasts are evaluated against, the conflict dyad at hand must have resulted
in at least 25 battle-related deaths over the course of the concerned calendar year. This criterion is not applied in the UCDP-Candidate data that informs the ViEWS forecasts on a monthly basis in addition
to the GED data, and has therefore been excluded from the outcome definitions above. When evaluating the forecasts against GED data, it is nevertheless implicitly applied. More about the UCDP-Candidate
dataset, and its differences from the GED data, can be found in the 2021 presentation article of the UCDP-Candidate dataset.
[2] In line with UCDP coding procedures, the government of a state is defined as the party controlling the capital of the state, whether or not the party is the de jure holder of power.
[3] An incompatibility is by the UCDP defined as a stated challenge over the governmental power or over a specified territory.
[4] Armed groups are here defined as any non-governmental group of people that have announced a name for their group and that uses.
[2] In line with UCDP coding procedures, the government of a state is defined as the party controlling the capital of the state, whether or not the party is the de jure holder of power.
[3] An incompatibility is by the UCDP defined as a stated challenge over the governmental power or over a specified territory.
[4] Armed groups are here defined as any non-governmental group of people that have announced a name for their group and that uses.
Levels of analysis
The VIEWS forecasts are presented at two levels of analysis: the country-month level and sub-national PRIO-GRID-month level.
Temporal unit: Calendar months
Geographic scope: Global (country level), Africa and the Middle East (grid level)
Geographic scope: Global (country level), Africa and the Middle East (grid level)
The country-month level
The country-month level uses the calendar month as the temporal unit of analysis and countries as the spatial unit. The set of countries is derived from the
Gleditsch & Ward (1999) list of independent states, while their geographic extent (and country IDs) are determined by the GIS dataset CShapes (
Weidmann, Kuse & Gleditsch, 2010).
Please note that the choice of country set and delimitations thereof was made on methodological grounds and does not reflect the views or opinions of the VIEWS team.
The PRIO-GRID-month level
The PRIO-GRID-month (pgm) level uses calendar months as the temporal units of analysis, and spatial units derived from PRIO-GRID 2.0 (
Tollefsen, Strand & Buhaug, 2012). The latter is a standardized spatial grid structure consisting of quadratic grid cells that jointly cover all areas of the world at a resolution
of 0.5 x 0.5 decimal degrees, approximately 55×55 km around the equator.
Data partitioning
In order to train and calibrate the forecasting models used in VIEWS, all available data are split into two sets of data partitions. One is used for true forecasting, and the other for evaluating historic forecasts.
The forecasting periodization is used for true/actual forecasting. The
data are split into four partitions: a training, calibration, “predictor updating”, and forecasting period.
- The
training period runs from the first month of available UCDP GED data, i.e. January 1990, up until the month before the start of the calibration period. The length of the training
period is increased by one year following the annual release of the UCDP GED dataset as we then re-train our models.
- The
calibration period is 48 months long. Its start and end dates shift by one year following the annual release of the UCDP GED data and our subsequent re-training of the VIEWS models. If the
most recent UCDP GED release covers the year of 2021 (and the VIEWS models have been retrained accordingly), the calibration period ends on 31 December 2021; if the last release covers the year
of 2022, it runs up until 31 December 2022; and so forth.
- The
predictor updating period runs from the month following the last month of available UCDP GED data up until and including the last month of available UCDP Candidate data. During
the predictor updating period, the VIEWS system is informed by the latter as a monthly substitute for the UCDP GED data, in addition to updates from other predictors that follow a regular update
schedule. The predictor updating period is thus extended by one month each time we generate and release a new set of monthly VIEWS forecasts.
- The
forecasting period is the rolling 36-month period for which we release true forecasts each month. The forecasting period starts immediately after the last month of available UCDP Candidate
data, i.e. after the last month of input data informing the VIEWS models. The name of each data release in the
VIEWS API reflects the last month of input data and can thus be used to deduce when the predictor updating period ends and the forecasting period starts. The fatalities002_2023_08_t01 release of VIEWS data, e.g., was informed by data up until and including Aug 2023, and thus contains forecasts for September 2023 – August 2026.
For the fatalities002_2023_08_t01 release of VIEWS data,
e.g., the partitions looked at follows:
- Training period: Jan 1990 – Dec 2017
- Calibration period: Jan 2018 – Dec 2021 (awaiting retraining of the models with the UCDP GED data for 2022)
- Predictor updating period: Jan 2022 – Aug 2023
-
Forecasting period: Sept 2023 – Aug 2026