Military named entity recognition (MNER) is one of the key technologies in military information extraction. Traditional methods for the MNER task rely on cumbersome feature engineering and specialized domain knowledge. In order to solve this problem, we propose a method employing a bidirectional long short-term memory (BiLSTM) neural network with a self-attention mechanism to identify the military entities automatically. We obtain distributed vector representations of the military corpus by unsupervised learning and the BiLSTM model combined with the self-attention mechanism is adopted to capture contextual information fully carried by the character vector sequence. The experimental results show that the self-attention mechanism can improve effectively the performance of MNER task. The F-score of the military documents and network military texts identification was 90.15% and 89.34%, respectively, which was better than other models.
Understanding the spatial and temporal variations of evapotranspiration (ET) is vital for water resources planning and management and drought monitoring. The development of a satellite remote sensing technique is described to provide insight into the estimation of ET at a regional scale. In this study, the Surface Energy Balance Algorithm for Land (SEBAL) was used to calculate the actual ET on a daily scale from Landsat-8 data and daily ground-based meteorological data in the upper reaches of Huaihe River on 20 November 2013, 16 April 2015 and 23 March 2018. In order to evaluate the performance of the SEBAL model, the daily SEBAL ET (ETSEBAL) was compared against the daily reference ET (ET0) from four theoretical methods: the Penman-Monteith (P-M), Irmak-Allen (I-A), the Turc, and Jensen-Haise (J-H) method, the ETMOD16 product from the MODerate Resolution Imaging Spectrometer (MOD16) and the ETVIC from Variable Infiltration Capacity Model (VIC). A linear regression equation and statistical indices were used to model performance evaluation. The results showed that the daily ETSEBAL correlated very well with the ET0, ETMOD16, and ETVIC, and bias between the ETSEBAL with them was less than 1.5%. In general, the SEBAL model could provide good estimations in daily ET over the study region. In addition, the spatial-temporal distribution of ETSEBAL was explored. The variation of ETSEBAL was significant in seasons with high values during the growth period of vegetation in March and April and low values in November. Spatially, the daily ETSEBAL values in the mountain area were much higher than those in the plain areas over the study region. The variability of ETSEBAL in this study area was positively correlated with elevation and negatively correlated with surface reflectance, which implies that elevation and surface reflectance are the important factors for predicting ET in this study area.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.