Abstract: The lost values are a common problem in many pivotal real application like data-mining machine learning and pattern detection algorithms. In now days ,there are huge streams of information which contain a missing values for many reasons for instance a malfunction in a piece of equipment, a tissue section on a slide was not stained properly,a technician forgot to enter a value in a spreadsheet, etc. If they are neglected, they will be loosed as well as, another information related. This research is suggested a new mathematical approach for solving the missing value problem based on a polynomial fitting model. The proposed model (polynomial curve fitting based on segmentation of variable point and variable modes SVPVM algorithm)is applied on PAMAP2 that contain 2872532 records each of them have 54 values. Also system is implemented on samples of dataset after some points are removed. Finally, the proposed approach is executed on known function such as sine wave. The results are compared with another methods for retrieval missing points such as linear and mean methods. In general the proposed model has been produced good results according toevaluation methods such as mean square error. So, the proposed method is expected to give promising results in the field of information loss.
Nabeel H. Al-A`araji, Eman S. Al-Shamery and Alyaa Abdul Hussein, 2016. A New Polynomial Curve Fiting Based on Segmentation of Variable Point and Variable Modes for Reconstructing Missing Values. Research Journal of Applied Sciences, 11: 1089-1094.