During my Phd studies on Polish Academy of Science I conducted researches associated with the use of AI (Artificial Intelligence) in medicine to analyse microarrays gene expression in case of fight with children cancer called as Neuroblastoma. During my studies on Military Academy School in Warsaw I acted with AI in robotics. But my first serious project was for Army and based on AI. Below I’d like to present solution based on mentioned technology for Money Fraud Detection – because in the past I dealt with similar solution for one of the biggest bank of the world.
In this part we focus on the assumptions and tools used to our project. Let’s start… Do You remember the main picture from the current article? If Not, then let’s look again to better understand whole logic used in present solution..
Our project is built on four main architectonic main blocks:
- Client Side (www.angular.io, https://facebook.github.io/react/)- for analysts and people to monitoring transactions and alerts, analysis. It will be based on React.js or Angular 2 to present results of data from AI evaluations
- Total.js / Node.js (www.totaljs.com)- REST API and Linkage/Coupling server for communication between the client, database and platform R (as a platform for artificial intelligence). Additional we will use node modules to RT communication: socket.io or primus.io node modules.
- DB – Database to storage our data used to analysis – we can use SQL or NoSql DB
- R Server (https://mran.microsoft.com, http://revolutionanalytics.com) – professional AI platform, old Deploy R Platform. Thanks to R Server Platform You can create powerful, portable analytics applications based on different analytics methods: Neural Networks (Linear/No Linear, Hopfiled Network), Genetic Algorithms, Artificial Immune Systems, Fuzzy Logic, Graph etc. In our solution we use open platform provided by Microsoft and accessible on https://mran.microsoft.com
Server R gives support for a full range of R-based analytics, big data statistics, predictive modeling, and machine-learning capabilities. When your data stores grow, Microsoft R Server can be deployed to perform at scale wherever your big data lives—including databases such as SQL Server 2016, Hadoop clusters, data warehouses, and even data stores in the cloud. You can build artificial intelligence-enabled applications based on R language by using machine learning and artificial intelligence.
R Server features mentioned by Microsoft:
Analyzing data wherever are storage – cloud, different distribution of Databases — without having to move it.
Building modules of artificial intelligence based on Your own algorithms or provided by Microsoft or another companies
- Easy process of deployment to a variety of platforms at scale and very robust security.
- Scale R analytics for big data – Scale analytics from individual servers to large clusters as your business needs change.
- Amazing speed of analysis a big data in cloud in real time thanks to a powerful computational powers – this what we are most interested
- R Server can run on different OS system platform: Microsoft, Linux, OS X
Main page R Server – http://revolutionanalytics.com
In every separate article we will focus on each from mentioned our blocks. Let’s look into companies related with our tools – big brands and Major players in the global software market: Microsoft, Node.js, Google, Facebook..
See You o the next article where we focus on technical issues 🙂