ARIMA-modellering av anlagda bilbränder i Sverige
(2019) STAH11 20182Department of Statistics
- Abstract
- An ongoing trend the last couple of years has been the intense reporting of an increased amount of deliberate vehicle fires in Sweden. As the overall assumption in society is that the number of deliberate vehicle fires is steadily increasing, a statistic model trying to investigate this was made. In 1998, the share of deliberate car fires were estimated to 12%, the same number in 2015 was 38%, indicating a growing problem. In addition, these types of criminal behaviour is characterized by a high number of criminals not getting convicted since there seldom is neither a lot of
witnesses nor technical evidence.
By dividing data collected from the Swedish Civil Contingencies Agency into three subcategories, namely large cities,... (More) - An ongoing trend the last couple of years has been the intense reporting of an increased amount of deliberate vehicle fires in Sweden. As the overall assumption in society is that the number of deliberate vehicle fires is steadily increasing, a statistic model trying to investigate this was made. In 1998, the share of deliberate car fires were estimated to 12%, the same number in 2015 was 38%, indicating a growing problem. In addition, these types of criminal behaviour is characterized by a high number of criminals not getting convicted since there seldom is neither a lot of
witnesses nor technical evidence.
By dividing data collected from the Swedish Civil Contingencies Agency into three subcategories, namely large cities, medium-sized towns and smaller towns, a time series analysis was conducted. To further concretize the difference between these subcategories, the data was collected as a relative number, the number of deliberate car fires per 1000 residents. All available data in the database of the Swedish Civil Contingencies Agency were used, meaning the analysis contained data between the years 1998 and 2017.
The analysis was made by fitting ARIMA-models, Autoregressive Integrated Moving Average-models, to the data in order to enable the creation of forecasts as well as prediction intervals. By removing a subset of the last 10% of the data before the analysis (datapoints from 2016 and 2017), a form of reference was made to observe the quality of the predictions. The prediction intervals were found to be too narrow in the case of large cities and smaller towns when compared to the subset. A possible reason for this may be that the real observations made between 2016 and 2017 contained sporadic unforseen events, such as rebellions. The feared positive upwarding trend was confirmed as all predictions predicted a steady increase in these types of crimes. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8972279
- author
- Persson, Amelie LU
- supervisor
-
- Jonas Wallin LU
- organization
- course
- STAH11 20182
- year
- 2019
- type
- M2 - Bachelor Degree
- subject
- language
- Swedish
- id
- 8972279
- date added to LUP
- 2019-03-25 11:39:44
- date last changed
- 2019-03-25 11:39:44
@misc{8972279,
abstract = {{An ongoing trend the last couple of years has been the intense reporting of an increased amount of deliberate vehicle fires in Sweden. As the overall assumption in society is that the number of deliberate vehicle fires is steadily increasing, a statistic model trying to investigate this was made. In 1998, the share of deliberate car fires were estimated to 12%, the same number in 2015 was 38%, indicating a growing problem. In addition, these types of criminal behaviour is characterized by a high number of criminals not getting convicted since there seldom is neither a lot of
witnesses nor technical evidence.
By dividing data collected from the Swedish Civil Contingencies Agency into three subcategories, namely large cities, medium-sized towns and smaller towns, a time series analysis was conducted. To further concretize the difference between these subcategories, the data was collected as a relative number, the number of deliberate car fires per 1000 residents. All available data in the database of the Swedish Civil Contingencies Agency were used, meaning the analysis contained data between the years 1998 and 2017.
The analysis was made by fitting ARIMA-models, Autoregressive Integrated Moving Average-models, to the data in order to enable the creation of forecasts as well as prediction intervals. By removing a subset of the last 10% of the data before the analysis (datapoints from 2016 and 2017), a form of reference was made to observe the quality of the predictions. The prediction intervals were found to be too narrow in the case of large cities and smaller towns when compared to the subset. A possible reason for this may be that the real observations made between 2016 and 2017 contained sporadic unforseen events, such as rebellions. The feared positive upwarding trend was confirmed as all predictions predicted a steady increase in these types of crimes.}},
author = {{Persson, Amelie}},
language = {{swe}},
note = {{Student Paper}},
title = {{ARIMA-modellering av anlagda bilbränder i Sverige}},
year = {{2019}},
}