The support vector machine(SVM)is a classical machine learning method.Both the hinge loss and least absolute shrinkage and selection operator(LASSO)penalty are usually used in traditional SVMs.However,the hinge loss i...The support vector machine(SVM)is a classical machine learning method.Both the hinge loss and least absolute shrinkage and selection operator(LASSO)penalty are usually used in traditional SVMs.However,the hinge loss is not differentiable,and the LASSO penalty does not have the Oracle property.In this paper,the huberized loss is combined with non-convex penalties to obtain a model that has the advantages of both the computational simplicity and the Oracle property,contributing to higher accuracy than traditional SVMs.It is experimentally demonstrated that the two non-convex huberized-SVM methods,smoothly clipped absolute deviation huberized-SVM(SCAD-HSVM)and minimax concave penalty huberized-SVM(MCP-HSVM),outperform the traditional SVM method in terms of the prediction accuracy and classifier performance.They are also superior in terms of variable selection,especially when there is a high linear correlation between the variables.When they are applied to the prediction of listed companies,the variables that can affect and predict financial distress are accurately filtered out.Among all the indicators,the indicators per share have the greatest influence while those of solvency have the weakest influence.Listed companies can assess the financial situation with the indicators screened by our algorithm and make an early warning of their possible financial distress in advance with higher precision.展开更多
The Chinese Area Positioning System （CAPS） is a positioning system based on satellite communication that is fundamentally different from the 3＂G＂ （GPS, GLONASS and GALILEO） systems. The latter use special-purpos...The Chinese Area Positioning System （CAPS） is a positioning system based on satellite communication that is fundamentally different from the 3＂G＂ （GPS, GLONASS and GALILEO） systems. The latter use special-purpose navigation satellites to broadcast navigation information generated on-board to users, while the CAPS transfers ground-generated navigation information to users via the communication satellite. In order to achieve accurate Positioning, Velocity and Time （PVT）, the CAPS employs the following strategies to over- come the three main obstacles caused by using the communication satellite： （a） by real-time following-up frequency stabilization to achieve stable frequency; （b） by using a single carrier in the transponder with 36 MHz band-width to gain sufficient power; （c） by incorporating Decommissioned Geostationary Orbit communication satellite （DGEO）, barometric pressure and Inclined Geostationary Orbit communication satellite （IGSO） to achieve the 3-D posi- tioning. Furthermore, the abundant transponders available on DGEO can be used to realize the large capacity of communication as well as the integrated navigation and communication. With the communication functions incorporated, five new functions appear in the CAPS： （1） combination of navigation and communication; （2） combination of navigation and high accu- racy orbit measurement; （3） combination of navigation message and wide/local area differen- tial processing; （4） combination of the switching of satellites, frequencies and codes; and （5） combination of the navigation message and the barometric altimetry. The CAPS is thereby labelled a PVT5C system of high accuracy. In order to validate the working principle and the performance of the CAPS, a trial system was established in the course of two years at a cost of about 20 million dollars. The trial constellation consists of two GEO satellites located at E87.5° and E110.5°, two DGEOs located at E130° and E142°, as well as barometric altimetry as a virtual satellite. Static and dynamic performance tests were completed for the Eastern, the Western, the Northern, the Southern and the Middle regions of China. The evaluation results are as follows： （1） land static test, plane accuracy range： C/A code, 15-25 m; P code, 5-10 meters; altitude accuracy range, 1- m; （2） land dynamic test, plane accuracy range, C/A code, 15-25 m; P code, 8-10m; （3） velocity accuracy, C/A code, 0.13-0.3 m s-1, P code, 0.15-0.17 m s- 1; （4） timing accuracy, C/A code, 160 ns, P code, 13 ns; （5） timing compared accuracy of Two Way Satellite Time and Frequency Transfer （TWSTFT）, average accuracy, 0.068 ns; （6） random error of the satellite ranging, 10.7 mm; （7） orbit determination accuracy, better than 2 m. The above stated random error is 1σ error. At present, this system is used as a preliminary operational system and a complete system with 3 GEO, 3 DGEO and 3 IGSO is being established.展开更多
Pseudorange bias has become a practical obstacle in the field of high-precision global navigation satellite system(GNSS)applications,which greatly restricts the further development of high-precision applications.Unfor...Pseudorange bias has become a practical obstacle in the field of high-precision global navigation satellite system(GNSS)applications,which greatly restricts the further development of high-precision applications.Unfortunately,no studies have been conducted on the pseudorange biases of the BeiDou navigation satellite system(BDS).To mitigate the effects of pseudorange biases on the BDS performance to the greatest extent possible,the origin of such BDS pseudorange biases are first thoroughly illustrated,based upon which the dependency of the biases on the receiver configurations are studied in detail.Owing to the limitations regarding the parameter re-settings for hardware receivers,software receiver technology was used to achieve the ergodicity of the receiver parameters,such as the correlator spacing and front-end bandwidth,using high-fidelity signal observations collected by a 40-m-high gain dish antenna at Haoping Observatory.Based on this,the pseudorange biases of the BDS B1I and B3I signals and their dependency on different correlator spacings and front-end bandwidths were adequately provided.Finally,herein,the suggested settings of the correlator spacing and front-end bandwidth for BDS receivers are in detail proposed for the first time.As a result,the pseudorange biases of the BDS signals will be less than 20 cm,reaching even under 10 cm,under this condition.This study will provide special attention to GNSS pseudorange biases,and will significantly promote a clear definition of the appropriate receiver parameter settings in the interface control documents of BDS and other individual satellite systems.展开更多
文摘The support vector machine(SVM)is a classical machine learning method.Both the hinge loss and least absolute shrinkage and selection operator(LASSO)penalty are usually used in traditional SVMs.However,the hinge loss is not differentiable,and the LASSO penalty does not have the Oracle property.In this paper,the huberized loss is combined with non-convex penalties to obtain a model that has the advantages of both the computational simplicity and the Oracle property,contributing to higher accuracy than traditional SVMs.It is experimentally demonstrated that the two non-convex huberized-SVM methods,smoothly clipped absolute deviation huberized-SVM(SCAD-HSVM)and minimax concave penalty huberized-SVM(MCP-HSVM),outperform the traditional SVM method in terms of the prediction accuracy and classifier performance.They are also superior in terms of variable selection,especially when there is a high linear correlation between the variables.When they are applied to the prediction of listed companies,the variables that can affect and predict financial distress are accurately filtered out.Among all the indicators,the indicators per share have the greatest influence while those of solvency have the weakest influence.Listed companies can assess the financial situation with the indicators screened by our algorithm and make an early warning of their possible financial distress in advance with higher precision.
基金supported by Ministry of science and Technology of China (Nos.2007CB815500, 2004AA105030and 2007AA12Z3423)Chinese Academy of science (No.KGCXI-21)the National Natural Science Foundation of China (No.10453001).
文摘The Chinese Area Positioning System （CAPS） is a positioning system based on satellite communication that is fundamentally different from the 3＂G＂ （GPS, GLONASS and GALILEO） systems. The latter use special-purpose navigation satellites to broadcast navigation information generated on-board to users, while the CAPS transfers ground-generated navigation information to users via the communication satellite. In order to achieve accurate Positioning, Velocity and Time （PVT）, the CAPS employs the following strategies to over- come the three main obstacles caused by using the communication satellite： （a） by real-time following-up frequency stabilization to achieve stable frequency; （b） by using a single carrier in the transponder with 36 MHz band-width to gain sufficient power; （c） by incorporating Decommissioned Geostationary Orbit communication satellite （DGEO）, barometric pressure and Inclined Geostationary Orbit communication satellite （IGSO） to achieve the 3-D posi- tioning. Furthermore, the abundant transponders available on DGEO can be used to realize the large capacity of communication as well as the integrated navigation and communication. With the communication functions incorporated, five new functions appear in the CAPS： （1） combination of navigation and communication; （2） combination of navigation and high accu- racy orbit measurement; （3） combination of navigation message and wide/local area differen- tial processing; （4） combination of the switching of satellites, frequencies and codes; and （5） combination of the navigation message and the barometric altimetry. The CAPS is thereby labelled a PVT5C system of high accuracy. In order to validate the working principle and the performance of the CAPS, a trial system was established in the course of two years at a cost of about 20 million dollars. The trial constellation consists of two GEO satellites located at E87.5° and E110.5°, two DGEOs located at E130° and E142°, as well as barometric altimetry as a virtual satellite. Static and dynamic performance tests were completed for the Eastern, the Western, the Northern, the Southern and the Middle regions of China. The evaluation results are as follows： （1） land static test, plane accuracy range： C/A code, 15-25 m; P code, 5-10 meters; altitude accuracy range, 1- m; （2） land dynamic test, plane accuracy range, C/A code, 15-25 m; P code, 8-10m; （3） velocity accuracy, C/A code, 0.13-0.3 m s-1, P code, 0.15-0.17 m s- 1; （4） timing accuracy, C/A code, 160 ns, P code, 13 ns; （5） timing compared accuracy of Two Way Satellite Time and Frequency Transfer （TWSTFT）, average accuracy, 0.068 ns; （6） random error of the satellite ranging, 10.7 mm; （7） orbit determination accuracy, better than 2 m. The above stated random error is 1σ error. At present, this system is used as a preliminary operational system and a complete system with 3 GEO, 3 DGEO and 3 IGSO is being established.
基金the National Nature Science Foundation of China(Nos.61501430 and 41604029)the State Key Laboratory of Geo-information Engineering(SKLGIE2017-M-2-2)。
文摘Pseudorange bias has become a practical obstacle in the field of high-precision global navigation satellite system(GNSS)applications,which greatly restricts the further development of high-precision applications.Unfortunately,no studies have been conducted on the pseudorange biases of the BeiDou navigation satellite system(BDS).To mitigate the effects of pseudorange biases on the BDS performance to the greatest extent possible,the origin of such BDS pseudorange biases are first thoroughly illustrated,based upon which the dependency of the biases on the receiver configurations are studied in detail.Owing to the limitations regarding the parameter re-settings for hardware receivers,software receiver technology was used to achieve the ergodicity of the receiver parameters,such as the correlator spacing and front-end bandwidth,using high-fidelity signal observations collected by a 40-m-high gain dish antenna at Haoping Observatory.Based on this,the pseudorange biases of the BDS B1I and B3I signals and their dependency on different correlator spacings and front-end bandwidths were adequately provided.Finally,herein,the suggested settings of the correlator spacing and front-end bandwidth for BDS receivers are in detail proposed for the first time.As a result,the pseudorange biases of the BDS signals will be less than 20 cm,reaching even under 10 cm,under this condition.This study will provide special attention to GNSS pseudorange biases,and will significantly promote a clear definition of the appropriate receiver parameter settings in the interface control documents of BDS and other individual satellite systems.
