Research On Feature Selection And Forecasting Method Of ALS Clinical Data

Amyotrophic lateral sclerosis(ALS)is a motor neuron disease that rare but often documented,and its symptoms include muscle weakness,paralysis and eventually death,usually within 3 to 5 years from disease onset.In order to help the clinical nursing and determine the new disease predictors,this paper mainly aims at the fea-ture selection and prediction methods of clinical data of ALS,and predict the pro-gression of ALS in the short term.The specific research work includes:(1)Feature selection method was adopted to select the characteristic dataset of ALS,and some of the most effective feature subsets are selected from the feature group of the patient to reduce the dimension space of the feature.What's more,whether the experimental dataset contains relevant information or redundant informa-tion is directly affect the later of classification performance.In this paper,we have adopted the random frog algorithm coupled with the partial least squares(RFA-PLS).Experiments show that the RFA-PLS algorithm has better performance than other feature selection methods in dealing with ALS datasets and effectively remove the high degree of correlation due to the combination of principal component analysis method,the selected features are representative as well.(2)For the selected subset of features that mentioned above,the partial least squares regression(PLSR)based on polynomial algorithm is adopted to predict the progression of ALS in the short term and used some of the more commonly regression algorithms to compare:support vector regression(SVR),and ridge regression(RR).The experimental results are evaluated by the root-mean-square deviation(RMSD)and Pearson correlation coefficient,indicate that the PLSR method has lower time complexity and the robustness is stronger than the other two methods.In conclusion,our method can estimate the future disease progression of ALS patients,is helpful to understand the ALS disease mechanisms,and play important role in making decisions regarding the test of the novel therapeutic approaches in clinical trials.