Coordinates are basic needs for both geospatial and non-geospatial professionals and as a result, geodesists have the responsibility to develop methods that are applicable and practicable for determining cartesian coo...Coordinates are basic needs for both geospatial and non-geospatial professionals and as a result, geodesists have the responsibility to develop methods that are applicable and practicable for determining cartesian coordinates either through transformation, conversion or prediction for the geo-scientific community. It is therefore necessary to implement mechanisms and systems that can be employed to predict coordinates in either two dimensional （2D） or three dimensional （3D） spaces. Artificial Intelligence （AI） techniques and conventional methods within the last decade have been proposed as an effective tool for modeling and forecasting in various scientific disciplines for solving majority of problems. The primary objective of this work is to compare the efficiency of artificial intelligence technique （Feed Forward Back propagation Neural Network （FFBPNN）） and conventional methods （Ordinary Least Squares （OLS）, General Least Squares （GLS）, and Total Least Squares （TLS）） in cartesian planimetric coordinate＇s prediction. In addition, a hybrid approach of conventional and artificial intelligence method thus, TLS-FFBPNN has been proposed in this study for 2D cartesian coordinates prediction. It was observed from the results obtained that FFBPNN performed significantly better than the conventional methods. However, the TLS-FFBPNN when compared with FFBPNN, OLS, GLS and TLS gave stronger and better performance and superior predictions. To further confirm the superiority of the TLS-FFBPNN the Bayesian Information Criterion was introduced. The BIC selected the TLS-FFBPNN as the optimum model for prediction.展开更多
We give a proof of an explicit formula for affine coodinates of points in the Sato’s infinite Grassmannian corresponding to tau-functions for the KdV hierarchy.
文摘Coordinates are basic needs for both geospatial and non-geospatial professionals and as a result, geodesists have the responsibility to develop methods that are applicable and practicable for determining cartesian coordinates either through transformation, conversion or prediction for the geo-scientific community. It is therefore necessary to implement mechanisms and systems that can be employed to predict coordinates in either two dimensional （2D） or three dimensional （3D） spaces. Artificial Intelligence （AI） techniques and conventional methods within the last decade have been proposed as an effective tool for modeling and forecasting in various scientific disciplines for solving majority of problems. The primary objective of this work is to compare the efficiency of artificial intelligence technique （Feed Forward Back propagation Neural Network （FFBPNN）） and conventional methods （Ordinary Least Squares （OLS）, General Least Squares （GLS）, and Total Least Squares （TLS）） in cartesian planimetric coordinate＇s prediction. In addition, a hybrid approach of conventional and artificial intelligence method thus, TLS-FFBPNN has been proposed in this study for 2D cartesian coordinates prediction. It was observed from the results obtained that FFBPNN performed significantly better than the conventional methods. However, the TLS-FFBPNN when compared with FFBPNN, OLS, GLS and TLS gave stronger and better performance and superior predictions. To further confirm the superiority of the TLS-FFBPNN the Bayesian Information Criterion was introduced. The BIC selected the TLS-FFBPNN as the optimum model for prediction.
文摘We give a proof of an explicit formula for affine coodinates of points in the Sato’s infinite Grassmannian corresponding to tau-functions for the KdV hierarchy.
