| |||||||||||||||||||||||||||
|
Customer Interviews: Cameron Cooper has been a instructor at Fort Lewis College in Southwestern Colorado for 4-years teaching developmental mathematics and computer science and also serves as an enrollment analyst. Mr. Cooper has an extensive educational background with a Bachelors in Mathematics from Occidental College, a Masters in Information Networking from Carnegie Mellon University, a Masters in Communications from Northwestern University and a Masters in Education from Harvard University. Mr. Cooper is a doctoral candidate in Applied Management & Decision Sciences at Walden University. Mr. Cooper has been using NeuroSolutions and NeuroSolutions for Excel for the past 2-years and has attended NeuroDimension's Neural Network Course. AT-RISK identifies "at risk" students for Developmental Mathematic courses, which have an average nationwide failure rates between 40% and 50% . AT-RISK uses various input factors such as learning styles, responses to questions about attitudes & beliefs regarding mathematics, high school GPA and standardized test scores to determine whether the student should take the support class designed around AT-RISK. The support class is 1-hour per week and it addresses frustration tolerance, test anxiety and much more. This class has increased the overall student success rate by 8% in the first semester compared to the previous year. Fort Lewis College has already adopted the AT-RISK neural network application for future semesters and is looking at implementing it for other problematic courses. In addition, Mr. Cooper has been approached by several other colleges to implement the AT-RISK project at their schools and he also has been approached by a national research organization for potential nationwide distribution. The AT-RISK project uses a standard Multilayer Perceptron (MLP), single hidden-layer network with approximately 50 weights. The initial network was created with 70 inputs, trimmed to 20 using Spearman's Rho Correlation and refined again to 8 inputs using logistic regression. The final network used 6 inputs which was finalized using backwards elimination. Mr. Cooper used NeuroSolutions for Excel's Leave-N-Out Training feature to create a model that was 81% accurate in predicting "at risk" students. Cameron had many great things to say about why he uses NeuroSolutions such as the "latitude in constructing neural networks" and that NeuroDimension "always incorporates the latest technologies" in NeuroSolutions. Mr. Cooper also described his experience with NeuroDimension's support and services as an "invaluable resource". We would like to thank Mr. Cooper for sharing his success with NeuroSolutions and we hope to continue to see great things from Mr. Cooper and AT-RISK in the near future! If you are interested in contacting Mr. Cooper about the AT-RISK project, please email him at cooper_c -at- fortlewis.edu. |
||||||||||||||||||||||||||
|