Develop a PHP script to evaluate quality loan applications (AI / neural network)

Your job is to develop a PHP script that evaluates the quality of loan application based on a given set of input parameters. The system's input is the loan application data in JSON format and the output must be an integer between 0 and 100, 0 meaning the loan application seems very low quality (i.e. applicant probably won't pay back) and 100 meaning a high quality application (i.e. applicant probably is paying back). The system must be based to neural network and/or to genetic algorithm or other similar analysis method that has been trained using sample input and output data (old loan applications whose quality we now know). The final delivery must also contain the mechanism for training the system with more data in the future. Attached is an SQL dump of a 100 rows of sample data for reference. The data contains the input data and the desired output of the system (integer). If I accept your bid, I will provide you with the full data set of about few thousand rows of data. The SQL dump can also be found at: [login to view URL] You are free to use any available pattern matching, artificial intelligence, neural network etc libraries that are available for PHP assuming the libraries are free to use. If you wish to use some commercial code, please state it in your bid, if the commercial code seems good and the price is reasonable, I can accept that kind of solution too. Do not offer solution in any other programming languages. Do not ask for any other data of the loan applications or applicants, the JSON contains all the data we have. With your bid, please give a brief overview of how would you implement this.