Predicting Dyslexia using Machine Learning:

There are a series of questions that need to be answered by the user, and on the basis of the results, scores are given for language, memory, speed, visual and audio skills. ● The objective of this model is to give an accurate prediction of the chances of the user having dyslexia, using the above scores as input.

Working of Model:

The purpose of this project is to predict the ‘Label’ i.e., chances of the applicant having dyslexia, if the values of ‘Language_vocab’, ‘Memory’, ‘Speed’, ‘Visual_discrimination’, ‘Audio_Discrimination’ and ‘Survey_Score’ are provided. In the file DyslexiaML, we have created five different models to find the one which is the best fit for the given dataset. The comparison has been made between DecisionTree, RandomForest, SVM, RandomForest with GridSearch and SVM with GridSearch, on the basis of error, precision and recall. On comparing all of these parameters, we found that RandomForest with GridSearch is the best fit for our dataset.

On the basis of our findings, we then created the final model using RandomForestClassifier with GridSearchCV in order to make the most accurate predictions, in DyslexiaML_final file.
This model was then tested in Test_Model on a new dataset to find the labels, which were then compared with the actual label values. After this check we found out that our model was able to make predictions for dyslexia with a 5.8% error rate.

At last, we've created a user-friendly file Input_test, wherein the user is given a brief introduction of the model, followed by which they can enter the values of ‘Language_vocab’, ‘Memory’, ‘Speed’, ‘Visual_discrimination’, ‘Audio_Discrimination’ and ‘Survey_Score’, followed by which the final model created is used to provide the result in a readable format.