PREDIKSI TINGKAT KRIMINALITAS MENGGUNAKAN LONG SHORT TERM MEMORY (LSTM)

SAUD, AGUSTINA ROSARIA and Supardi, Julian (2025) PREDIKSI TINGKAT KRIMINALITAS MENGGUNAKAN LONG SHORT TERM MEMORY (LSTM). Undergraduate thesis, Sriwijaya University.

	Text RAMA_55201_09021082025173.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_55201_09021082025173_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_55201_09021082025173_0010077210_01_fornt_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (1MB)
	Text RAMA_55201_09021082025173_0010077210_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (359kB) | Request a copy
	Text RAMA_55201_09021082025173_0010077210_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (392kB) | Request a copy
	Text RAMA_55201_09021082025173_0010077210_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (1MB) | Request a copy
	Text RAMA_55201_09021082025173_0010077210_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (504kB) | Request a copy
	Text RAMA_55201_09021082025173_0010077210_06.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (107kB) | Request a copy
	Text RAMA_55201_09021082025173_0010077210_07_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (128kB) | Request a copy
	Text RAMA_55201_09021082025173_0010077210_08_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (5kB) | Request a copy

Abstract

Prediction or also called forecasting is the process of estimating events that will occur in the future. In this research, software is developed that can predict crime rates using the Long Short-term Memory method using the Los Angles crime dataset which has a crime type domain of 139 with details of 9 types of crimes that have the most frequency, namely; vehicle - stolen, battery - simple assault, theft from vehicle, identity theft, vandalism - serious crime, theft, assault with a deadly weapon, aggravated assault, ordinary theft - minor, intimate partner - ordinary assault, and theft from motor vehicles. The dataset has 962431 rows of data with a total of 20 parameters including; DR_NO, DATE OCC, TIME OCC, AREA, AREA NAME, Rpt Dist No, Crm Cd, Crm Cd Desc, Mocodes, Vict Age, Vict Sex, Vict Descent, Premise Desc, Weapon Used Cd, Weapon Desc, Status, Status Desc, LOCATION, LAT and LON. In this research, the model training is done with the configuration: split train validation data into 80%-20%, for 2 model architectures: 2 layer LSTM and 3 layer LSTM each trained with 50, 100, and 150 epochs to analyze time series data and predict crime rates more accurately. The Long Short Term Memory (LSTM) based crime prediction model has been successfully developed, which has the lowest prediction error among other configurations with an RMSE of 0.0208.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Prediction, Crime, Long Short Term Memory
Subjects:	Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation.
Divisions:	09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User:	Agustina Rosaria Saud
Date Deposited:	22 Jan 2025 03:58
Last Modified:	22 Jan 2025 03:58
URI:	http://repository.unsri.ac.id/id/eprint/166191

Actions (login required)

View Item