Impartial Intrusion & Crime Detection Without Gender or Caste Discrimination

Download Project Document/Synopsis

Data mining is supported by Automated information gathering systems for fraud and crime detection. Undoubtedly, banks, finance companies, vast organizations, insurance agencies, club, and so on are progressively mining information about their clients or representatives in perspective of distinguishing from potential interruption. In many organization intruders are identified by their outer look or by other sensitive attributes like gender, race or religion. This will not help to detect correct intruder. Detecting the intruder based on the sensitive attributes will not help to detect the correct intruder as well as it will lead to discrimination of particular person. Identifying the intruder by sensitive attributes is wrong method to detect the potential intruder. Our proposed system will detect the objective misbehavior of the potential intruder, rather than from sensitive attributes like gender, race or religion. Legal data will be stored in database based on these data intruder will be detected rather than sensitive attributes. Legal data includes behavior of past crime done by the intruder. If system identifies any suspected intruder, system will mine the data in database. These data will be compared with the behavior of the suspected person. If these data matches than system will declare the person to be potential intruder or criminal. Therefore, legitimate classification rules can still be extracted but discriminating rules based on sensitive attributes cannot.



Advantages
  • This system can be used in many government or private organizations where security plays a major role.
  • This system will be helpful to reduce fraud or crime.
  • This system can be used in many banking sectors for security purpose.
Disadvantages
  • As the system suspects only those intruder whose pattern matches with the past criminal. If the intruder uses new pattern system might not suspect the intruder.

Leave a Comment

Your email address will not be published.