Artificial Intelligence as a subject of computer science

• About this group
• Forums

• Home
• Groups
• Artificial Intelligence as a subje ...
• Forums
• General Chat
• Neuro-Fuzzy Modeling Techniques Applied to Short Essay Auto-Grading Problem

• 
  Dr. Yousef Arikat

  The company name is only visible to registered members.

  Neuro-Fuzzy Modeling Techniques Applied to Short Essay Auto-Grading Problem

  Abstract: Automated Essay Grading (AEG) systems that are currently available use different techniques to extract specific written dimensions features to assess the written prose. We have applied several Neuro-Fuzzy approaches to try to solve the short essay AEG problem. The main idea behind our approach is to identify the number of main keywords ( 5 input) each of which has 4 synonyms based on specific constraints. These inputs have been processed by developing two general models including; Artificial Neural Network Back-propagation optimization technique and Subtractive Clustering technique. Furthermore we have developed a third general model to incorporate more main keywords by processing double the number of previous models input, single output with specific constraints. We have combined Subtractive clustering and Neuro Fuzzy techniques. The Sub-Clustering Neuro Fuzzy model could be used for a variable number of inputs.
  900 generated datasets have been used representing ten questions related to information technology field. Datasets have been splited into training and testing sets using cross-validation method. We have used the correlation coefficient to measure the agreements between the actual and predicted marks of our models, and used two error measures including; Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE) to indicate the performance of the developed models.
  The average results obtained from all three general models whether using 5 or 10 main keywords (our datasets and IICM data) shows a correlation measure of 0.9977, while MAPE values are less than 0.001, where as RMSE are less than 0.053. Furthermore we have tested other system datasets on our developed model and shared our datasets with other systems ( Institute for Information systems and Computer Media (IICM), Graz University of Technology, Austria ).
  This indicates the suitability and adequacy of our developed models to solve the short essay AEG problem and shows that our preliminary results are promising. Further testing on different types of questions is required so as to conclude the suitability having achieved these promising results.
  
  • 15 Jun 2010, 10:00 am
  Dr. Yousef Arikat wrote: > Abstract: Automated Essay Grading (AEG) systems that are currently > available use different techniques to extract specific written > dimensions features to assess the written prose. We have applied > several Neuro-Fuzzy approaches to try to solve the short essay AEG > problem. The main idea behind our approach is to identify the number > of main keywords ( 5 input) each of which has 4 synonyms based on > specific constraints. These inputs have been processed by developing > two general models including; Artificial Neural Network > Back-propagation optimization technique and Subtractive Clustering > technique. Furthermore we have developed a third general model to > incorporate more main keywords by processing double the number of > previous models input, single output with specific constraints. We > have combined Subtractive clustering and Neuro Fuzzy techniques. > The Sub-Clustering Neuro Fuzzy model could be used for a variable > number of inputs. > 900 generated datasets have been used representing ten questions > related to information technology field. Datasets have been splited > into training and testing sets using cross-validation method. We have > used the correlation coefficient to measure the agreements between > the actual and predicted marks of our models, and used two error > measures including; Mean Absolute Percentage Error (MAPE) and Root > Mean Square Error (RMSE) to indicate the performance of the developed > models. > The average results obtained from all three general models > whether using 5 or 10 main keywords (our datasets and IICM data) > shows a correlation measure of 0.9977, while MAPE values are less > than 0.001, where as RMSE are less than 0.053. Furthermore we have > tested other system datasets on our developed model and shared our > datasets with other systems ( Institute for Information systems and > Computer Media (IICM), Graz University of Technology, Austria ). > This indicates the suitability and adequacy of our developed > models to solve the short essay AEG problem and shows that our > preliminary results are promising. Further testing on different types > of questions is required so as to conclude the suitability having > achieved these promising results.