Using Convolutional Neural Networks to Build a Lightweight Flood Height Prediction Model with Grad-Cam for the Selection of Key Grid Cells in Radar Echo Maps

Recent climate change has brought extremely heavy rains and widescale flooding to many areas around the globe. However, previous flood prediction methods usually require a lot of computation to obtain the prediction results and impose a heavy burden on the unit cost of the prediction. This paper proposes the use of a deep learning model (DLM) to overcome these problems. We alleviated the high computational overhead of this approach by developing a novel framework for the construction of lightweight DLMs. The proposed scheme involves training a convolutional neural network (CNN) by using a radar echo map in conjunction with historical flood records at target sites and using Grad-Cam to extract key grid cells from these maps (representing regions with the greatest impact on flooding) for use as inputs in another DLM. Finally, we used real radar echo maps of five locations and the flood heights record to verify the validity of the method proposed in this paper. The experimental results show that our proposed lightweight model can achieve similar or even better prediction accuracy at all locations with only about 5~15% of the operation time and about 30~35% of the memory space of the CNN.

LSALO-BP: A Hybrid Model for Children Adulthood Height Prediction

The prediction of children's adult height is a common procedure in childhood endocrinology. Through the prediction of children's adult height, it is possible to find abnormalities in children's growth and development. Many jobs in today's society have certain requirements for height, so the accuracy of children adulthood height prediction is important for children. Current methods for predicting adult height of children have some shortcomings such as inaccurate accuracy. To deal with these problems, this paper analyzes the data collected by the Chinese children and adolescents' physical and growth health projects in primary and secondary schools in Zhejiang Province, and proposes a method for predicting adult height based on back propagation neural network (BPNN) with the body composition of children and adolescents as input. Since the BP algorithm has the risk of falling into local optimization, and we propose LSALO-BP model that incorporates the ant lion optimizer (LSALO) into the BP algorithm as location strategy to avoid local optimization. The improvements achieved by the ant lion algorithm are mainly reflected in: improving the ant's walk mode, and enhancing the global search ability of the LSALO algorithm. The comparison experiment of 10 benchmark functions proves the feasibility and effectiveness of the location strategy. The LSALO-BP model is applied to the prediction of adult height of children and adolescents. The experimental results show that compared with other models, the LSALO-BP prediction model has increased the prediction accuracy by 6.67%~16.08% for boys and 4.67%~6.6% for girls, which can more accurately predict the adult height of children and adolescents.

PREDICTING MAXIMAL COUNTERMOVEMENT JUMP HEIGHT FROM UPRIGHT AND SQUAT POSITIONS HEAD TITLE: UPRIGHT AND SQUAT MAXIMUM JUMP HEIGHT PREDICTORS

Introduction. Countermovement jump is common in sport and testing and performed from various starting positions. Little is known about effective contributors to maximal countermovement jump height from various starting positions. Purpose and Objectives. Determine effective jump height predictors and effect of starting position on countermovement jump height. Applied Methodology. Forty-nine collegiate athletes performed maximal height countermovement jumps from upright and squatting positions with arm movement. Several variables were calculated from kinetic data. Correlation and regression determined variables related to and predictive of jump height in both conditions. Paired t-tests evaluated differences in jump height. Achieved Major Results. Upright condition jump height positively correlated with peak force and power, eccentric and concentric impulses, and countermovement depth. Jump height prediction included peak force and power, and eccentric and concentric impulses. Squat condition jump height positively correlated with peak force and power, mean rate of force development, force generated at the beginning of propulsion, and concentric impulse. Jump height prediction equation included mean rate of force development, force at the beginning of propulsion, and peak power. Jump height was higher in the upright condition. Conclusions. Higher jumps are achieved from the upright position. Peak force, peak power, and concentric and eccentric impulses best contribute to upright jump height. Mean rate of force development, force at the beginning of propulsion, and peak power best predicted squat jump height. Limitations. We did not restrict arm movement, to encourage natural motion. Depth was not controlled, rather advising a comfortable depth. Subjects were recruited from various collegiate sports. Practical implications. Maximal jump height from various positions may be achieved through efforts to maximize jump peak power and increase musculotendinous loading in sport-specific starting positions. Originality/Value. This is the first study to explore the predictors of upright and squat countermovement jumps. These results can guide jump performance training.

Bead Geometry Prediction in Laser-Wire Additive Manufacturing Process Using Machine Learning: Case of Study

In Laser Wire Additive Manufacturing (LWAM), the final geometry is produced using the layer-by-layer deposition (beads principle). To achieve good geometrical accuracy in the final product, proper implementation of the bead geometry is essential. For this reason, the paper focuses on this process and proposes a layer geometry (width and height) prediction model to improve deposition accuracy. More specifically, a machine learning regression algorithm is applied on several experimental data to predict the bead geometry across layers. Furthermore, a neural network-based approach was used to study the influence of different deposition parameters, namely laser power, wire-feed rate and travel speed on bead geometry. To validate the effectiveness of the proposed approach, a test split validation strategy was applied to train and validate the machine learning models. The results show a particular evolutionary trend and confirm that the process parameters have a direct influence on the bead geometry, and so, too, on the final part. Several deposition parameters have been found to obtain an accurate prediction model with low errors and good layer deposition. Finally, this study indicates that the machine learning approach can efficiently be used to predict the bead geometry and could help later in designing a proper controller in the LWAM process.

Height Prediction Based on the Lengths of Ulna and Tibia in an Iranian Population

Background: Height is an important factor for medical, nutrition, and forensic sciences; nevertheless, measuring proves to be a challenging task in some cases. In this respect, an alternative immediate, accurate, and possible anthropometric evaluation is needed. Objectives: This study was set to find a proper formula to estimate height from the lengths of the ulna and tibia in Iranian adults. Methods: A total of 500 healthy males and females aged 20-40 years were randomly selected from the volunteers’ pool for this cross-sectional study. Ulna and tibia lengths and standing heights were measured according to standard protocols. Ulna and tibia lengths were applied to find a reliable equation to predict stature accurately. Data were statistically analyzed by SPSS version 17 using regression, curve estimation, and linear model. Results: The mean (SD) heights of male and female participants were 176.45 (11.98) cm and 161.29 (10.11) cm, respectively, while the right and left ulna and the right and left tibia were 29.05 (1.63) cm, 29.03 (1.44) cm, 38.86 (1.33) cm, and 38.88 (1.25) cm, respectively. The correlation coefficients of r = 0.80 and r = 0.69 for males and females participants’ right ulna, respectively, showed a significant correlation with height. Hence, the new formula provided reliable results for stature estimation for northern Iran subgroups. Conclusions: Equations based based on right ulna length are more reliable and accurate for height prediction in both genders. It should be considered that these equations could be different among ethnically diverse populations, even in the northern Iranian population.

Forecasting of Typhoon-Induced Wind-Wave by Using Convolutional Deep Learning on Fused Data of Remote Sensing and Ground Measurements

Taiwan is an island, and its economic activities are primarily dependent on maritime transport and international trade. However, Taiwan is also located in the region of typhoon development in the Northwestern Pacific Basin. Thus, it frequently receives strong winds and large waves brought by typhoons, which pose a considerable threat to port operations. To determine the real-time status of winds and waves brought by typhoons near the coasts of major ports in Taiwan, this study developed models for predicting the wind speed and wave height near the coasts of ports during typhoon periods. The forecasting horizons range from 1 to 6 h. In this study, the gated recurrent unit (GRU) neural networks and convolutional neural networks (CNNs) were combined and adopted to formulate the typhoon-induced wind and wave height prediction models. This work designed two wind speed prediction models (WIND-1 and WIND-2) and four wave height prediction models (WAVE-1 to WAVE-4), which are based on the WIND-1 and WIND-2 model outcomes. The Longdong and Liuqiu Buoys were the experiment locations. The observatory data from the ground stations and buoys, as well as radar reflectivity images, were adopted. The results indicated that, first, WIND-2 has a superior wind speed prediction performance to WIND-1, where WIND-2 can be used to identify the temporal and spatial changes in wind speeds using ground station data and reflectivity images. Second, WAVE-4 has the optimal wave height prediction performance, followed by WAVE-3, WAVE-2, and WAVE-1. The results of WAVE-4 revealed using the designed models with in-situ and reflectivity data directly yielded optimal predictions of the wind-based wave heights. Overall, the results indicated that the presented combination models were able to extract the spatial image features using multiple convolutional and pooling layers and provide useful information from time-series data using the GRU memory cell units. Overall, the presented models could exhibit promising results.