Dosimetria

The Intelligent Systems are characterized, or the ability to use knowledge to solve problems, or the ability to make associations and inferences of complex problems from small tasks to real problems.

This tool aims, aims to help the process of dosimetry pen, Branch Criminal Law, in accordance with current legislation, taking into account the current doctrine and jurisprudence.

This application uses a neural network (Backpropagation) to calculate the Phase 1 dosimetry in criminal law.

The database used for training was based on another system using Fuzzy Logic (end of course work Computer Engineering UnP 2015).

The proposed tool in this work is intended to aid decision-making by the judge at the time of establishing the penalty base of a defendant, which means focusing only on the first step, the three set out in art. 68 CP (Penal Code), to determine the sentence to be served.

This clarification is necessary in view of the subjective nature of the analysis of the legal circumstances, for the judge at the time have to motivate your decision based on the eight circumstantial aspects of crime. therefore the tool presented here provides an aid to this process, with the valuation of the circumstances and the consequent impact on the increase or decrease pen-based.

They represent the first phase of the dosimetry pen, are also called unnamed circumstances, given their exhaustive reiteration by law, shall be borne by the discretion of the judge analysis, but may not represent a reduction below the minimum penalty provided by law, or, as little beyond the maximum.

The input parameters are: Culpability, Background, Social Conduct, Agent Personality, Crime Reason, Circumstance Crime, Consequence of Crime and behavior of the victim.

It has been tested empirically with random values ​​of the input variables, considering the defendant sentenced for the simple crime of murder (worth 6-20 years).

 - Culpability - P1 (0 to 10) - Low - High.

 - Background - P2 (0 to 10) - Good background - bad history.

 - Social conduct - P3 (0 to 10) - Good behavior - bad conduct

 - Personality - P4 (-5 to 5) - Pacato - Normal- violent

 - Reason - P5 (-5 to 5) - Acceptable - Fair - reprehensible

 - Consequences - P6 (0 to 10) - Take - Record

 - Victim's behavior - P7 (0 to 10) - not helped - caused

 - Circumstance - P8 (0 to 10) - Irrelevant - barbarian

It is worth noting the differentiation in personality variables and reason, which are determined -5 input values ​​5. This differentiation exists solely to demonstrate that levels near zero, there is to consider increase or decrease the penalty , being these circumstances reaching the level of the average man, for which there is greater or smaller fail.

It is important to note that the variable "victim behavior", unlike

too, the lowest values ​​input will result in a more severe apenação the defendant, while higher values ​​will represent a lesser culpability of the defendant, should be his sentence base more softened.

The application usage can be summarized as follows:

1. The user enters the defaults using the Defaults button. By default, the application has 15 patterns with their respective desired value. However, the user can change these values.

2. Enter the parameters of RNA.

3. Click the Train button.

4. the calculation of the base sentence.