AI concepts are heling the users in making the automatic decision based on the fed instruction s and algorithms to the system. In this article the author has explained the implementation process of the AI and other expert system in decision making process and also in creating a neural network. In interruption recognition terms, abuse location is for coordinating known assaults; and irregularity location is for finding obscure assaults. The flow explore in both interruption recognition and spam discovery are on abnormality recognition (semi-regulated) and unaided methodologies. In interruption recognition examine, the utilization of grouping to decrease information and Gee for inconsistency identification had been famous.With the available clustered data base system one have to analyse and have to create a network for this system so that they can get the nearest and the most suitable data set which will help them in getting the most nearest and suitable solution. This research proposal consists the overall analysis of the of the financial fraud detection with the help of Big data and other machine learning techniques.The visuals and representations of the data in the simple forms of graphs and charts which helped in analysing processes is also elaborated in this research paper.
In the process of data mining and data analysis complex data are simplified in the normal form for the purpose of analysis and representation. In this article review the researcher is going to research and analyse the different types of articles related to the process of data mining and their implementation of their techniques for the purpose of data mining in the in financial fraud detection. There are various kinds of data mining techniques which can be implemented or can be used in these process for simplifying the cluster of data.
Article 1:- The application of data mining techniques in financial fraud detection: A classification framework and an academic review of literature.
This article consists the detailed analysis and review about the implementation of the data mining and big data in financial fraud detection. In this article the learner is going to the in depth analysis of different case studies and methodologies related to the detection process and the data mining process. Frauds and other theft related to the data and other things is getting very common. With every coming day users are getting different types of threats related to their data and information. Above all these they are also getting threats related to financial thefts and thugs. In the recent days there are a lot of case registered which directly relates to the financial fraud. In the year 2014 -15 more than 400 financial frauds case has been registered in a local police station of the Queensland, Australia. Most of the cases were related to the online theft and frauds. In most of the cases which are detected by the, local police they have found that these frauds were related to the frauds and other theft from the context of the data mining (Okeke,2018).
In the process of fraud detector with the data mining a number of key words and other theyare used in this process. They related and attached these keywords with the available data bases to compare the existing records of the victims and the criminals. IN this process lot of newoutcomes are generated which helped the organization in getting new results and otheroutcomes. From the processes of data mining they detected or overviewed the outcome of the whole database.
This article contains the detailed analysisof the fraud detection with the help of data mining. There they have explained the number of mining algorithm which are used in the detecting process and also in data mining or data warehousing process. The author of the article has explained in details about the implementation of different types of algorithm which will help them in analysing and mining the data in proper manner. There are different algorithms like Bayesian algorithms which allows the data process in creating the risk network for this process.
Highlighting more on the concept of the Bayesian algorithm which the author has discussed in this article, they are using this for generating the new risk table from the clustered data. They are processing these clustered data as a trainee data and creating new risk table and using as test data. With all these test data and the trainee data users can get the outcomes and profound related to the big data analysis (Roiger,2018).
In the application of data mining in fraud detection there are mainly two types of data which can be found, on these supervised data which has the direct relation with the lists and data of the frauds cases and other one is unsupervised data where the usershave to process these data to their desired results and outcomes related to the fraud detection. In this article they have found that both of these datatypes are helping them in different ways .In the case of the unsupervised data they are able to create a new matrix with the same data for the purpose of creating different matrices and tables (Abdallah,2016).
Discussing about the existing fraud detection system, according to Williams et. al.(1997) with using one of the most used algorithm in the process of the data analysis and other technologies algorithm “KNN algorithm”. From the KNN (K nearest algorithm) they can implement the process of “hot spots methodology”. In this methodology users finds the most nearest and the most suitable data from the test data set. In this article the author has explained this things with an example. In this example the author has explained a case study of the bank fraud which took place in the year 2004. In this case a cyber-criminal withdrawn the amount of $ 2million form the different accounts of the same bank. And was using these amounts in the transaction process for and was transferring process. The police and the local cybercrime experts collected the data of the transaction location and matched withal the available data set. It provided them the accessibility to get the set of IP address which the attacker was using to perform the transaction. With the help of these IP they traced all of them one by one and finally they were able to catch that person. After sometime the bank faced the same issue of fraud and this time they matched the data with the.
With all these explained case studies and implementation it can be concluded that the process of data mining with applying the different types of algorithms and paradigms can helping many efficient ways in solving the financial cases which the organizations and individuals are facing. Here the author has also explained about the future scope of these data mining processes which will provide them the easiness to solve the cases in future. In the last section the author has explained about the some recommendation which they have to follow. They can keep the record of the access and the transaction process and are advised to match with the available data set. This step will help the banks in verifying their customers and account holders also in the case of the transaction process they need to implement the same system, they can check the frauds details and the location to verify whether the customers who is accessing this account is either getting matched with any of the risk table which the banks have or not. If they are getting a match then the process will be automatically aborted unless it will proceed as a normal process.
In the end of this article the author has concluded with the simple statement of Bayesian belief network, which says that in the process of data mining and data ware housing there is nothing like the perfect outcome. With the available clustered data base system one have to analyse and have to create a network for this system so that they can get the nearest and the most suitable data set which will help them in getting the most nearest and suitable solution.
In interruption recognition terms, abuse location is for coordinating known assaults; and irregularity location is for finding obscure assaults. The flow explore in both interruption recognition and spam discovery are on abnormality recognition (semi-regulated) and unaided methodologies. In interruption recognition examine, the utilization of grouping to decrease information and Gee for inconsistency identification had been famous. detail that k-intends to pack information and report that Well performed somewhat better than occurrence based learning (IBL) for semi-genuine client level information. Thus use SOM to diminish information for Well displaying. The creator demonstrate that various Well models with fluffy rationale can be utilized to lessen false positive rates. Additionally, advocate Scanty Markov Transducers. Be that as it may, reason that straightforward static approaches, for example, event recurrence conveyances and cross entropy between appropriations, outflank Well. Other irregularity discovery thinks about what's more, Most investigations finish up that inconsistency recognition doesn't execute just as abuse identification. Solo methodologies incorporatewhat's more,which supporter replicator neural systems to recognize exceptions. To recognize spam from the email server, encodes email as hash-based content, and stays away from marks using report space thickness to discover enormous volumes of comparative messages. On a little example of marked messages, the creator’s remark that SVM is the best regulated calculation in any case, the location time is unreasonably long for an occasion driven framework. The utilization of game hypothesis to display the key cooperation between the framework and enemy has been as of late brought into interruption and spam identification inquire about. apply hypothetical game hypothesis to represent the collaboration between one assailant and one hub to distinguish interruptions from versatile impromptu systems. In spam discovery, adjust game hypothesis to consequently re-get familiar with a cost-touchy regulated calculation given the cost-touchy enemy's ideal technique. It characterizes the enemy and classifier ideal technique by making some legitimate suppositions. Tried under various false positives costs, the game-theoretic guileless Bayes classifier beats the traditional classifier by proficiently predicting its positives and negatives.
In the process of decisions making and other methodologies one can implement the concept of the data mining and other processes. In this article the author has discussed the implantation process of decision tree and other concepts in data mining process which will be furtherimplemented for the useof fraud detection and other activities related to the detection of crimes and thefts.
Decision trees are machine learning concepts which are implemented as a key attribute in creating the decisions and other concepts.in this article the author has highlighted the concept of the decision making process with thehelp of the big data and machinelearning concepts.
Initiating from the normal programming concept of IF THEN and ELSE which is widely used in most of the available programming languages. The author has highlighted the implementation of this concept in the big data process and machine learning for creating the suitable decision with the most appropriate solutions. Pointing out the process of decision making the author has discussed about the process of decision making step by step. In his context the author has explained that from the available clustered data they can create some simple sets of data and then can create the decision from all these sets first andthen will analyse all the generated sets with this process.
For the implementation in the fraud detection system for the financial cases the author has also explained one of the case study where they sued this decisions making process in solving the cases related to the financial fraud. In this case study the author has discussed the case of United Bank fraud case where their own employees have created a trouble for their company. In this process they have used the decision making, text mining process to solve the situation or the case.
With the involvement of artificial networks in latest technologies and work processes they are making the tasks simpler than before, AI concepts are heling the users in making the automatic decision based on the fed instruction s and algorithms to the system. In this article the author has explained the implementation process of the AI and other expert system in decision making process and also in creating a neural network. For this research work where the researches is conducting the research of the fraud detection system in financial fields, this will relate in many different possible ways and can provide the suitable solution for detecting these cases and solving the case. Here in this article the author has explained about the system related to the data mining and data analysis related to the big data concepts (Pierson, 2016).
Artificial networks represents the strong mathematical expressions with the different type so of the mathematical operators like summations, exponents and other parameters which creates the network for the cases.
The implementation of the AI neural network will help the analyses process in creating the dataset with different variables and parameters. Here the researcher has practically examine a data set form the article to show the process of data mining and to implement them in the crimes and frauddetection system. The researcher has performed the application on a clustered data to sub set the data in different samples or in the form of the test data. With the outcome of this process the organizations will be able to verify the defaulters and the fraud list.
Continuing this research work for the process of fraud detection system for the big data and other data mining processes. The researcherhas analysed this article which was based on the fraud detection and is implemented with the concepts of the data analysis and big data theory in respect to the time complexity and other futuristic parameters.
According to the author of the article in the traditional methods of fraud detestation in financial processes the departments and the related organizations were using the SQL methods, which was taking too much time and was unable to deliver the perfect solution for data mining and other process. In the SQL method it was a tough task to analyse all the data with the help of SQL queries, alsofor analyse these clustered data for the public information they need go choose for the analysis.
In this article the author has taken a case study of the insurance company. In this case study the author has explained about the process which an insurance company has selected to detect the defaulters and the frauds lists in their company, in the earlier methodologies which they were following was related with the SQL injection and query process. With this process they analysed the data and got the result but the data which they got was not sufficient and neither was relevant to the actual data which they were looking for. To avoid these things the company migrated to the big data and data mining process. Wot the shifting of the technology they got the unexpected result, it helped them in simplifying their data set in efficient manner. Data related to their customers like the healthcare reports and the codes are separated in separate table which provided them an initial helping tracing and monitoring the data. In the other process of keeping the records of the outdated data the insurance company was also able to filter the data easily.
In the context of data ware housing the author has also discusses about the process of future scopes of these systems. For the organization who are relate with the insurance process and are managing the financial data of the different business firms they can implement the big data concept which will help them in performing the data analysis in simpler manner. The process which the insurance company has selected for this assignment has a wide future scope in improving their service quality for their customers. The way which they have followed to manage their clustered data collection was quiet efficient in making the process easier and handing the processes in simple way.
At the end of the article the author has concluded with the result of the implementation process which the insurance comply has used. The company got the separate table and well-structured database of all their data in a tabular form. Above all these things they got the visuals and representations of the data in the simple forms of graphs and charts which helped them in analysing processes.