Intro to CSCI-130 Computing: Science & Applications (NS)
Layered Architecture LAYEROrder Application SW : Excel & Access 2 High-order P.L. : Visual Basic 1 Low-order P.L. : Assembly 3 System Software : O.S. 3 Machine Language 4 Data Representation 5 HW: Circuit Design6
General vs. Special Computers Computers can either be Special-purpose computers (Majority) Hardwired to do specific tasks only (usually one) i.e. execute one program Ubiquitous --- we interact with them almost daily --- embedded Examples? General-purpose computers Provide means to change their programs thus becoming multi- or general-purpose machines Include desktops, laptops/notebooks, servers, etc… People tend to associate the word “computer” only with them Limit ourselves to the latter type only
Software Software is the set of all programs that run on a computer VS Hardware : “Hard” Has a physical presence Comes in two forms System Software: controls the computer Applications Software: accomplishes user-defined tasks
Programs & Algorithms Characteristics of an algorithm : List of steps to complete a task Each step is PRECISELY defined and is suitable for the machine used Increase the value of X Jump! Add 5 to variable X The process terminates in a finite amount of time No infinite loops Written in an English-like language (Pseudocode)
Programs & Algorithms Program: A formal representation of a method for performing some task Written in a programming language understood by a computer Detailed and very well-organized (computers just do what they are told) Follows an algorithm … method for fulfilling the task Plan to do something VS the actual performance
Course Theme Consumer Credit Risk Prediction: the process of estimating the risk of loss due to a customer's non re-payment (default) on a consumer credit product, such as a mortgage, unsecured personal loan, credit card, overdraft etc... Problem : Given information for a new credit applicant, predict whether to approve or deny credit
Sample Data Customer Credit score ( ) Total income ($) Duration for present employment (years) Age (years) # of dependents Prediction (Class Label) A80061, YES B71072, YES C42029, NO D39022, NO E55088, NO F825101, YES G58943, NO ….
The k-NN Prediction Algorithm If something walks like a duck, quacks like a duck, looks like a duck, it must be a duck! In other words, find the “k” closest customers to the new applicant and use majority voting to predict the class label for that customer