Search

Density peaks clustering algorithm (DP) has difficulty in clustering large-scale data, because it requires the distance matrix to compute the density and $\delta $ -distance for each object, which has $O(n^{2})$ time complexity. Granular ball (GB) is a coarse-grained representation of data. It is based on the fact that an object and its local neighbors have similar distribution and they have high possibility of belonging to the same class. It has been introduced into supervised learning by Xia et al. to improve the efficiency of supervised learning, such as support vector machine, $k$ -nearest neighbor classification, rough set, etc. Inspired by the idea of GB, we introduce it into unsupervised learning for the first time and propose a GB-based DP algorithm, called GB-DP. First, it generates GBs from the original data with an unsupervised partitioning method. Then, it defines the density of GBs, instead of the density of objects, according to the centers, radius, and distances between its members and centers, without setting any parameters. After that, it computes the distance between the centers of GBs as the distance between GBs and defines the $\delta $ -distance of GBs. Finally, it uses GBs’ density and $\delta $ -distance to plot the decision graph, employs DP algorithm to cluster them, and expands the clustering result to the original data. Since there is no need to calculate the distance between any two objects and the number of GBs is far less than the scale of a data, it greatly reduces the running time of DP algorithm. By comparing with $k$ -means, ball $k$ -means, DP, DPC-KNN-PCA, FastDPeak, and DLORE-DP, GB-DP can get similar or even better clustering results in much less running time without setting any parameters. The source code is available at https://github.com/DongdongCheng/GB-DP.

Due to simplicity, K-means has become a widely used clustering method. However, its clustering result is seriously affected by the initial centers and the allocation strategy makes it hard to identify manifold clusters. Many improved K-means are proposed to accelerate it and improve the quality of initialize cluster centers, but few researchers pay attention to the shortcoming of K-means in discovering arbitrary-shaped clusters. Using graph distance (GD) to measure the dissimilarity between objects is a good way to solve this problem, but computing the GD is time-consuming. Inspired by the idea that granular ball uses a ball to represent the local data, we select representatives from a local neighborhood, called natural density peaks (NDPs). On the basis of NDPs, we propose a novel K-means algorithm for identifying arbitrary-shaped clusters, called NDP-Kmeans. It defines neighbor-based distance between NDPs and takes advantage of the neighbor-based distance to compute the GD between NDPs. Afterward, an improved K-means with high-quality initial centers and GD is used to cluster NDPs. Finally, each remaining object is assigned according to its representative. The experimental results show that our algorithms can not only recognize spherical clusters but also manifold clusters. Therefore, NDP-Kmeans has more advantages in detecting arbitrary-shaped clusters than other excellent algorithms.

ABSTRACTRecently, Convolutional Neural Network (CNN) has achieved some success in synthetic aperture radar (SAR) image classification. This outstanding performance mainly depends on a large number of training samples, but achieved satisfactory classification results with limited training samples remains a challenge. To address this problem, we propose a dual-scale Siamese densely connected network with Markov Random Fields (DS-SDCNet-MRF) for single-polarization SAR image classification. First, the Siamese densely connected network (SDCNet) is proposed to fully extract discriminative features under limited samples. Then, the proposed DS-SDCNet is constructed with two SDCNets of different scales to produce complementary classification results. Among them, the large-scale SDCNet has a better classification result in the homogeneous region, while the small-scale SDCNet tends to provide good detail preservation. Finally, an improved MRF model which combined the category probability information is proposed to further improve the classification performance. Experimental results on simulated and real single-polarization SAR data demonstrate that the proposed method achieves more encouraging classification performance than the current state-of-the-art classification methods with limited samples.

The application of controlled-release fertilizer was an effective measure to improve the fertilizers utilization, increase yield, and reduce environmental pollution. This study aimed to explore the effects of application times and frequency and fertilization ratio of controlled-release compound fertilizer with different nutrient release periods on apple production. The experimental design consisted of controlled-release compound fertilizers with nutrient release period of 3 months (C3), controlled-release compound fertilizers with nutrient release period of 6 months (C6), and common compound fertilizer (CCF). The treatments included (1) one-time application of C6 in March (B-C6); (2) one-time application of C6 in October (H-C6); (3) one-time application of 50% C3 and 50% C6 in March (B-(50C3 + 50C6)); (4) one-time application of 70% C3 and 30% C6 in March (B-(70C3 + 30C6)); (5) common compound fertilizer (CCF) with one-time application of CCF in March (B-CCF); and (6) common compound fertilizer with three-time application of CCF in March, June, and October (CT). The index of apple yield, apple tree growth index, fruit quality, and soil nutrient status were evaluated in 2-year filed production system. The results indicated that the treatment of B-(50C3 + 50C6) significantly improved the growth of apple trees on trunk diameter, plant height, new shoot growth and chlorophyll content, and the apple yield increased by 8.82%. By comprehensive comparison of yield, soil nutrients, and fruit tree growth indexes, the treatment with (B-(50C3 + 50C6)) was considered to be the best controlled-release fertilizer application method.

ABSTRACTRecently, convolutional neural networks (CNNs) have been widely used for synthetic aperture radar (SAR) image classification because of their powerful feature extraction ability and high performance. However, extracting discriminative features with limited training samples is still a challenge. Moreover, some samples may be image edge samples, which often contain multiple image categories, thus deteriorate classification accuracy. To address these issues, we propose a novel classification framework, named dense-connected global covariance network (DGCNet) with edge sample constraint (ESC). First, a dense-connected sub-network was designed, which can connect different convolutional layers of conventional CNN to strengthen feature propagation, encourage feature reuse, and alleviate gradient vanishing problem. Then, a global covariance pooling layer was introduced to fully exploit the second-order information of deep features and reduce the number of training parameters. Finally, an ESC strategy was integrated into DGCNet to further improve the classification performance by assigning a smaller weight to edge samples than non-edge samples during the training process. Experimental results on two datasets demonstrated that the proposed method achieves better classification results than several popular classification methods with limited training samples.

Although solid-phase activation of lignite using a nanocatalyst has great potential in producing low-cost and sustainable humic acid, the large-scale application of this technology still faces challenges because of the high price and toxicity of the nanocatalyst. Additionally, the specific molecular components of humic acid in activated lignite remain unknown. In this work, a multifunctional molybdate-phosphorus hierarchical hollow nanosphere (Mo-P-HH) catalyst was successfully manufactured by a simple way followed by phosphorization. In comparison with a commercial Pd/C catalyst, the multifunctional Mo-P-HH catalyst was more effective in producing water-soluble humic acid with small molecular functional groups from lignite via solid-phase activation. Moreover, Fourier transform ion cyclotron resonance mass spectrometry revealed the molecular compositions of humic acid in activated lignite. Compared with that from raw lignite, the humic acid after Mo-P-HH activation had less aromatic structure but higher content of lipids, proteins, amino sugar, and carbohydrates. In addition, the activated humic acid simulated seed germination and seedling growth. Therefore, this study provided a high-performance hierarchical hollow nanocatalyst for activation of humic acid and also offered the theoretical basis for the application of humic acid in agriculture.

The complex synthesis process, low utilization, and single function of fungicides have seriously hindered the development of fungicides in resistance to rice sheath blight. Here, an inexpensive and multifunctional Cu(II)-based water-dispersible humic acid (Cu-WH) fungicide with growth-promoting ability was developed with a simple method. A 3D molybdate carbon hierarchical nanosphere (MoO2-C-HN) catalyst was successfully synthesized using a green route and applied in a solid-phase activation of lignite to obtain water-dispersible humic acid. Cu(II)-based water-dispersible humic acid (Cu-WH) was then formed through a simple reaction of Cu(II) and the humic acid. The resultant Cu-WH showed strong antifungal performance against Rhizoctonia solaniin laboratory incubation experiments. After being treated with Cu3-WH (0.1 mg L–1), the control efficiency of rice sheath blight at 1, 3, and 5 days after infection was 90.54%, 78.96%, and 66.31%, respectively. It also enhanced the water-holding capacity of the substrate and thus effectively improved the growth of rice seedlings. In comparison to commercial rice seedling substrate, the substrate treated with 8 wt % of Cu3-WH increased plant height, stem diameter, fresh weight, and chlorophyll content by 19.23%, 35.91%, 14.52%, and 42.85%, respectively. The newly developed Cu-WH thus can be used as a novel low-cost efficient fungicide and growth stimulator to treat rice sheath blight as well as to increase rice production.

Although commercialized slow-release fertilizers coated with petrochemical polymers have revolutionarily promoted agricultural production, more research should be devoted to developing superhydrophobic biopolymer coatings with superb slow-release ability from sustainable and ecofriendly biomaterials. To inform the development of the superhydrophobic biopolymer-coated slow-release fertilizers (SBSF), the slow-release mechanism of SBSF needs to be clarified. Here, the SBSF with superior slow-release performance, water tolerance, and good feasibility for large-scale production was self-assembly fabricated using a simple, solvent-free process. The superhydrophobic surfaces of SBSF with uniformly dispersed Fe3O4superhydrophobic magnetic-sensitive nanoparticles (SMNs) were self-assembly constructed with the spontaneous migration of Fe3O4SMNs toward the outermost surface of the liquid coating materials (i.e., pig fat based polyol and polymethylene polyphenylene isocyanate in a mass ratio 1.2:1) in a magnetic field during the reaction-curing process. The results revealed that SBSF showed longer slow-release longevity (more than 100 days) than those of unmodified biopolymer-coated slow-release fertilizers and excellent durable properties under various external environment conditions. The governing slow-release mechanism of SBSF was clarified by directly observing the atmosphere cushion on the superhydrophobic biopolymer coating using the synchrotron radiation-based X-ray phase-contrast imaging technique. Liquid water only contacts the top of the bulges of the solid surface (10.9%), and air pockets are trapped underneath the liquid (89.1%). The atmosphere cushion allows the slow diffusion of water vapor into the internal urea core of SBSF, which can decrease the nutrient release and enhance the slow-release ability. This self-assembly synthesis of SBSF through the magnetic interaction provides a strategy to fabricate not only ecofriendly biobased slow-release fertilizers but also other superhydrophobic materials for various applications.

It is critical to evaluate the quality of clusters for most cluster analysis. A number of cluster validity indexes have been proposed, such as the Silhouette and Davies–Bouldin indexes. However, these validity indexes cannot be used to process clusters with arbitrary shapes. Some researchers employ graph-based distance to cluster nonspherical data sets, but the computation of graph-based distances between all pairs of points in a data set is time-consuming. A potential solution is to select some representative points. Inspired by this idea, we propose a novel Local Cores-based Cluster Validity (LCCV) index to improve the performance of Silhouette index. Local cores, with local maximum density, are selected as representative points. Since graph-based distance is used to evaluate the dissimilarity between local cores, the LCCV index is effective for obtaining the optimal cluster number for data sets containing clusters with arbitrary shapes. Moreover, a hierarchical clustering algorithm based on the LCCV index is proposed. The experimental results on synthetic and real data sets indicate that the new index outperforms existing ones.

ABSTRACTOsteosarcoma (OS) is one of the most common primary bone tumors and has a high disablity rate and case-fatality rate. The protracted stagnancy of the chemotherapy program and surgical technology for OS treatment prompted us to focus on the mechanisms of cancer carcinogenesis progression in OS.Nucleoside diphosphate kinase B (NME2) is a type of nucleoside diphosphate kinase that plays an important role in cellular processes. In this study, we report overexpression of NME2 in OS cell lines and correlate this overexpression with the clinicopathologic features of osteosarcoma. We used si-NME2 to downregulate expression of NME2 in OS cell lines. The results of the CCK8 and clone forming assays show that NME2 promotes OS cell line proliferation. Western blot assays show that deregulation of NME2 results in enhanced the expression of c-Myc, which promotes OS proliferation.

To reduce the preparation cost of superabsorbent and improve the N release rate at the same time, a novel low-cost superabsorbent (SA) with the function of N slow release was prepared by chemical synthesis with neutralized acrylic acid (AA), urea, potassium persulfate (KPS), and N,N’-methylenebis(acrylamide) (MBA). The order of influence factors on the water absorbency property was determined by an orthogonal L18(3)7experiment. On the basis of the optimization results of the orthogonal experiment, the effects of a single factor on the water absorption were investigated, and the highest water absorbency (909 g/g) was achieved for the conditions of 1.0 mol urea/mol AA ratio, 100% of AA neutralized, K+, 1.5% KPS to AA mass fraction, 0.02% MBA to AA mass fraction, 45 °C reaction temperature, and 4.0 h reaction time. The optimal sample was characterized by scanning electron microscopy (SEM) and Fourier transform infrared spectroscopy (FTIR). Swelling behaviors of the superabsorbent were investigated in distilled water and various soil and salt solutions. The water-release kinetics of SA in different negative pressures and soils were systematically investigated. Additionally, the maize seed germination in various types of soil with different amounts of SA was proposed, and the N could release 3.71% after being incubated in distilled water for 40 days. After 192 h, the relative water content of SA-treated sandy loam, loam, and paddy soil were 42, 56, and 45%, respectively. All of the results in this work showed that SA had good water retention and slow N-release properties, which are expected to have potential applications in sustainable modern agriculture.

Graphene-based mixed-dimensional materials hybridization is important for a myriad of applications. However, conventional manufacturing techniques face critical challenges in producing arbitrary geometries with programmable features, continuous interior networks, and multimaterials homogeneity. Here we propose a generalized three-dimensional (3D) printing methodology for graphene aerogels and graphene-based mixed-dimensional (2D + nD, where nis 0, 1, or 2) hybrid aerogels with complex architectures, by the development of hybrid inks and printing schemes to enable mix-dimensional hybrids printability, overcoming the limitations of multicomponents inhomogeneity and harsh post-treatments for additives removal. Importantly, nonplanar designed geometries are also demonstrated by shape-conformable printing on curved surfaces. We further demonstrate the 3D-printed hybrid aerogels as ultrathick electrodes in a symmetric compression tolerant microsupercapacitor, exhibiting quasi-proportionally enhanced areal capacitances at high levels of mass loading. The excellent performance is attributed to the sufficient ion- and electron-transport paths provided by the 3D-printed highly interconnected networks. The encouraging finding indicates tremendous potentials for practical energy storage applications. As a proof of concept, this general strategy provides avenues for various next-generation complex-shaped hybrid architectures from microscale to macroscale, for example, seawater desalination devices, electromagnetic shielding systems, and so forth.

Reasonably utilize the recyclable waste cardboard (175 million tons every year) had attracted the increasing environmental concern. However, the poor hydrophobicity and the loose porous structure of the cardboard-derived coating materials caused the rapid release of nutrients, thereby limiting the development and application of the bio-derived synthetic polymer coated controlled-release fertilizers (BPCFs). In this work, the hydrophobic-densified double-modified waste-carton-derived controlled-release fertilizers (HDCFs) were developed with nontoxic modifying agent and the simple production technology. The controlled release abilities of HDCFs were significantly enhanced (< 2 h to 120.45 days) and the nutrient release prediction models were established. The enhanced performance was attributed to the improved hydrophobicity and the obviously compact coating structure characterized by the three-dimensional computerized tomography (5.76–1.08 %). Furthermore, the enhanced elasticity (5025.52–1325.68 MPa) of the HDCFs coatings also contributed to improve the controlled-release abilities. The controlled-release mechanism was also clarified: the atmosphere “stopper” in the “smaller and less” micropores in HDCFs coating only allows water vapor molecules (instead of liquid) slowly permeate into the internal urea core and significantly enhance the controlled-release longevities. The dramatically increased oilseed rape yield (71.75 %) showed the efficient application effect of the HDCFs. All the results indicate that HDCF with 90:10 of the proportion of the castor oil and liquefaction polyhydric alcohols from cardboard (LPAC) and 5 % of the siloxane of the total polyols exhibits the best performance effect. This work provides the efficient strategy to foster the end-user confidence in the low-cost and eco-friendly biowaste-derived controlled-release fertilizers.

Instance reduction is aimed at reducing prohibitive computational costs and the storage space for instance-based learning. The most frequently used methods include the condensation and edition approaches. Condensation method removes the patterns far from the decision boundary and do not contribute to better classification accuracy, while edition method removes noisy patterns to improve the classification accuracy. In this paper, a new hybrid algorithm called instance reduction algorithm based on natural neighbor and nearest enemy is presented. At first, an edition algorithm is proposed to filter noisy patterns and smooth the class boundaries by using natural neighbor. The main advantage of the algorithm is that it does not require any user-defined parameters. Then, using a new condensation method based on nearest enemy to reduce instances far from decision line. Through this algorithm, interior instances are discarded. Experiments show that the hybrid approach effectively reduces the number of instances while achieves higher classification accuracy along with competitive algorithms.

The Prins condensation of isobutylene with formaldehyde over scillica-supported silicotungstic acid catalysts has been investigated. The catalysts were characterized via XRD, FTIR, Py-IR, NH3-TPD and TG–DSC techniques. It was found that the catalyst with 5 wt% silicotungstic acids loading exhibited the best isobutene conversion and isoprene selectivity, and excessive silicotungstic acid was not conducive to improve the catalytic performance. It was proposed that the concentration and strength of acid sites could affect the catalytic properties. An appropriate amount of weak Lewis acid sites were the active site for the condensation of isobutene with formaldehyde, and the medium-strong acid sites could cause the side reactions and coke deposits.

Bio-based polyurethanes are promising for the controlled release of nutrients and fertilizers, but their toughness and plasticity need to be improved. We developed a smooth, dense, elastic, and indestructible bio-based polyurethane (BPU) coating with a nutrient controlled release ∼150% superior, a tensile strength ∼300% higher, and a toughness ∼1200% higher than those for the original BPU coating. Through a one-step reaction of soybean oil polyols (accounting for more than 60%), isocyanate, and benzil dioxime, the dynamic covalent network based on oxime-carbamate replaces part of irreversible covalent cross-linking. The dynamic fracture-bonding reaction in the modified coating BPU can effectively promote the hydrogen bond recombination and oxime-carbamate chain migration in the coating process, which avoids the structural defects caused by coating tear and fertilizer collision. This work provides a simple and versatile strategy for building controlled-release fertilizer coatings.