77results about How to "Guaranteed retention" patented technology

The invention belongs to the technical field of intelligent agricultural data analysis, and relates to a crop identification and growth process prediction method and system based on multi-modal data. The method comprises the following steps: collecting growth process images and multi-modal environment data of crops in each growth stage; constructing a crop feature data set; constructing a sample data set; training to obtain a crop type identification model; obtaining a crop type identification result; generating a fusion feature vector; performing data enhancement processing; constructing a long-short-term memory network for training to obtain a growth stage prediction model; and predicting the growth stage of the crops. Crop type identification is performed by constructing a deep neural network model, feature fusion is performed on crop growth process image features and multi-modal environment data, crop growth stage prediction is performed by using a long-short-term memory network, fertilization, irrigation and pest control measures are optimized in combination with the multi-modal environment data, and the crop growth stage prediction accuracy is improved. The crop yield and quality are improved.

The invention discloses a cooking robot control method and system based on a knowledge graph, and relates to the technical field of cooking robot control, and the method comprises the steps: collecting multi-source cooking data, constructing the knowledge graph, receiving user demands, recommending dishes based on the graph, determining cooking parameters according to the recommended dishes, and controlling the cooking robot according to the cooking parameters. Comprise time, temperature, seasoning dosage and the like. The parameters are sent to a cooking robot controller, data are collected by a sensor during cooking, the data are compared with standard data, and the parameters are adjusted by a feedback algorithm to ensure that cooking meets the expectation. According to the method, personalized dish recommendation and cooking parameter adaptive adjustment can be realized, dish quality and taste are ensured, user health is considered, nutrition balance is promoted, the intelligent degree of the cooking robot is improved, the market competitiveness of the cooking robot is enhanced, and the requirement of people for high-quality cooking experience is met.

The invention relates to the technical field of photovoltaic panel fault diagnosis, and discloses a photovoltaic panel fault diagnosis method based on multi-source data fusion, and the method comprises the steps: collecting and marking data, collecting the data from an operation monitoring system and a field detection instrument of a photovoltaic panel, and carrying out the data expansion, and training the feature extraction model, the feature dimension reduction model and the classifier model in sequence, and finally inputting the collected data into the models for diagnosis. According to the method, the sample quality is enhanced by collecting and labeling photovoltaic panel operation data, performing data expansion and de-entanglement processing, feature extraction is performed by adopting a neural network based on adaptive fluctuation optimization, feature dimension reduction is realized by combining a self-encoding neural network of manifold mapping learning, and the classification precision is enhanced through a high-order neural network. According to the method, the accuracy and generalization ability of fault diagnosis are remarkably improved, the stability and efficiency of model training are optimized, a complex data mode is effectively dealt with, and stable operation of the photovoltaic panel in a variable environment is ensured.

The invention discloses a human body behavior recognition method and system based on space-time decoupling multi-scale feature fusion. The system comprises a local space-time multi-scale feature fusion network MDCNN and a space-time decoupling network DTST. Extracting different scale features of multichannel sensor data by using a multi-scale convolutional neural network, extracting shallow layer features and deep layer features of signals at the same time, and extracting features representing action information by fully utilizing complementary information among different level features; the method comprises the following steps of: decoupling a network into a time decoupling structure and a space decoupling structure through a time-space decoupling network based on Transform, so as to accurately extract time-space characteristics in an inertial signal; and finally, using a Softmax classifier to carry out action classification and recognition. According to the method, consistency and complementarity among multi-sensor data can be efficiently utilized, multi-scale spatial-temporal characteristics are fused, changes in time dimensions and relevant information in spatial positions are fully considered, different attention degrees are kept for signals of different sensing channels, and the recognition precision in human behavior recognition is improved.

The invention belongs to the technical field of medical data processing, and discloses a temporal-mandibular joint disease automatic diagnosis method and system based on artificial intelligence, and the method comprises the steps: collecting and preprocessing data; expanding the patient data, dynamically adjusting the structure and parameters of a data generator, and generating the patient data; performing feature extraction on the patient data, and automatically adjusting an evolution rule of a feature extraction network to obtain patient data features; performing dimensionality reduction on the data features to obtain low-dimensional patient data features; inputting the low-dimensional patient data features into a classifier, and dynamically adjusting the configuration and weight of hidden layer nodes of the classifier according to the low-dimensional patient data features to obtain a classification result; and defining a loss function, and training the model. Parameters can be automatically adjusted in the data feature extraction process, the data dimension reduction process is optimized by adopting quantum state mapping and an adaptive adjustment mechanism, the classifier can dynamically adjust the number of hidden layer nodes and an activation function, and the accuracy and robustness of a diagnosis result are improved.

The invention discloses an automobile software defect prediction method based on dynamic feature selection and oversampling, which comprises the following steps: acquiring a data set of automobile software defects, and defining a full feature set of the data set; performing data preprocessing on the data set, and dividing the data set into a training set and a test set; dynamically evaluating the contribution of each feature in the full feature set to a decision feature through a granularity decision entropy algorithm, and reserving important features with high information gain; new samples are generated for the minority class samples by introducing an intelligent resampling algorithm of auxiliary loss and diversity loss, and the representativeness and diversity of the minority class samples are enhanced; training the defect prediction model through the training set and testing the test set; and outputting a corresponding defect prediction result based on the characteristic data of the automobile software through the trained defect prediction model. According to the method, the accuracy of automobile software defect prediction and the generalization ability of the model in new data or multi-class and mixed type data environments can be improved.

The invention discloses a DAS-VSP data coherent noise suppression method and system based on adaptive texture analysis, and the method comprises the steps: respectively calculating vertical and horizontal derivatives and frequency values of effective signals and noise in synthesized DAS-VSP data, calculating the correlation between pixels, drawing a spatial correlation graph, and guiding the selection of a training strategy; expanding a traditional blind spot network into a neighborhood shielding network, rotating input data to construct a plurality of sub-graphs, and performing translation cutting to form a blind neighborhood; a corresponding blind neighborhood shielding strategy is designed, so that the network meets the condition that blind spot network noise pixels are independent; new asymmetric expansion convolution is proposed to ignore the influence of coherent noise; ordinary total variation regularization is expanded into texture total variation regularization, and texture weight factors are calculated according to different texture characteristics of signals and noise to smooth noise reserved signals; and performing a denoising test on the synthetic seismic data and the actual DAS-VSP seismic data, and analyzing a denoising effect through a quantitative index signal-to-noise ratio, structural similarity and a visualization effect.

The invention discloses a ghost imaging target reconstruction method based on FSD-GIRA, and relates to the technical field of digital image processing. According to the technical scheme, a fuzzy clustering algorithm is used for classifying feature values of an optical field intensity measurement matrix of an optical path detector, the feature values are reconstructed through singular value decomposition, dimensionality reduction processing of measurement data of a ghost imaging system is achieved, and the ghost imaging target reconstruction method based on FSD-GIRA is obtained. A WE-FOMP-CG algorithm is proposed, and recovery of sparse signals is optimized by combining generalized orthogonal matching pursuit, a conjugate gradient method and a multi-threshold wavelet noise reduction method; and selecting the most relevant sensing matrix array in each iteration, solving a linear equation set by using PCG, carrying out wavelet preprocessing and multi-threshold de-noising on a result, and updating and recovering estimation. The calculation efficiency and the convergence speed are improved, and a more accurate recovery result is provided in a high-noise environment. The invention also provides a WBDH algorithm, and the algorithm employs a dual denoising mechanism, including wavelet transform and a BM3D method, so as to guarantee that the restored image details and structures are reserved, and further improve the reconstruction quality of the ghost imaging target.

The invention provides an electrical equipment fault diagnosis method and device based on artificial intelligence, and relates to the technical field of fault recognition, and the method comprises the steps: obtaining historical operation data collected by a sensor installed in a power system, and enabling the historical operation data to serve as an original data set; expanding the original data set; inputting the expanded original data set into a feature extraction model to train the feature extraction model; training a feature dimension reduction model, and training a classifier model; and inputting real-time operation data collected by a sensor installed in the power system into the trained feature extraction model, the feature dimension reduction model and the classifier model in sequence, and outputting corresponding fault categories. The diversity of a data set is enhanced through a data expansion technology, the generalization ability of a model is improved, feature extraction, feature dimension reduction and classification processing are carried out, deep analysis and processing of data are achieved, and therefore the operation state and the possible fault type of the electrical equipment can be judged more accurately.

The invention relates to the technical field of intelligent manufacturing systems, and discloses an optimal scheduling search method for a multi-robot coordinated cooperation system, which comprises the following steps of: constructing a library timing Petri net model of the multi-robot coordinated cooperation system according to a robot resource use condition and a production process procedure of the multi-robot coordinated cooperation system; constructing a state diagram under a steepest excitation strategy based on a library time-assigned Petri net model, wherein the state diagram retains an optimal state sequence corresponding to the shortest processing time; designing a state merging rule in combination with the structure information of the Petri network, removing redundant states which do not belong to the optimal state sequence, and reducing the scale of the state diagram; the optimal scheduling scheme of the system can be obtained by searching the state diagram through global search. According to the method, a timing Petri net model which effectively represents a multi-robot resource conflict relationship and a processing event logic and time relationship is constructed, and a state space containing an optimal processing event sequence is constructed and searched on the basis, so that an optimal scheduling strategy is solved.

The invention discloses a surface treatment method for improving corrosion resistance of 304 copper-containing antibacterial stainless steel. The surface treatment method comprises the steps that 1, the surface of a 304 copper-containing antibacterial stainless steel plate to be treated is polished, cleaned and dried; 2, the surface of the plate is subjected to sand blasting treatment, and vibration cleaning is conducted after sand blasting is completed; 3, high-speed static friction treatment is conducted on the surface of the plate; 4, top-layer oxide skin in a static friction influence area on the surface of the stainless steel plate is removed, and light grinding and polishing are conducted; step 5, carrying out final surface cleaning and drying treatment on the 304 copper-containing antibacterial stainless steel plate subjected to surface treatment; according to the method for conducting sand blasting and static friction treatment on the surface of the commercial 304 copper-containing stainless steel, the grain size of the surface of the stainless steel is reduced, the structure uniformity of the stainless steel is improved, the service life of the stainless steel in a corrosive environment is prolonged, the surface treatment method is simple in process, efficient and free of pollution, and the corrosion resistance of the material is obviously improved.

The invention provides a remote sensing image desert shrub extraction method, which is characterized in that an FETNet model is obtained through improvement based on an OfficientNet-B0 baseline model, an original framework predefined in the baseline model is reserved at an encoder part, an EdgeViT module is introduced at a jump connection part of an encoder and a decoder, the global perception capability of the model is enhanced, and the algorithm is simple and easy to implement. An FEM module is introduced into a decoder part to ensure the reservation of space detail information, so that a finally generated desert shrub segmentation result is more accurate and meticulous, a preprocessed remote sensing image is input into an FETNet model, a stochastic gradient descent optimizer, a cosine annealing strategy and a cross entropy loss function are used for model training, the optimal weight of the FETNet model is obtained, and the final desert shrub segmentation result is obtained. The FETNet model is compared with a conventional model, and under the same test condition, all evaluation indexes in the remote sensing image desert shrub segmentation task obtain optimal results.

The invention discloses a zero-day malicious code detection model and method and a model construction device and method, and belongs to the technical field of malicious code detection. According to the model construction method, TransXNet serves as a basic model, training optimization is conducted on the basic model through a training set, then the zero-day malicious code detection model is obtained, diffusion operation is conducted on a visual malicious code image through a diffusion model DDIM, a new malicious code image is generated, and therefore the zero-day malicious code detection model is obtained. And combining the new malicious code image with the original malicious code image to jointly form training set data. By the adoption of the construction method, the detection accuracy and robustness of the model for the zero-day malicious code are improved, the problem of data set imbalance is solved, and therefore an effective technical scheme is provided for solving the problems existing in existing zero-day malicious code detection.

The invention discloses an AI computer vision target detection method based on an adaptive compensation mechanism. The method comprises the following steps of data acquisition and labeling, data expansion, feature extraction model training, feature dimension reduction model training, classifier model training and target detection and recognition. Compared with the prior art, the method has the advantages that 1, the expression of morphological characteristics is enhanced through the generative adversarial network, the quality and details of the generated image are improved, and the training effect and generalization ability of the model are enhanced; 2, the exploration type gradient descent method optimizes the training process, and by predicting and adjusting a gradient updating strategy, the training time is shortened, and the stability and performance of the model are improved; 3, the reverse reconstruction strategy of the auto-encoder effectively ensures the retention of key features, and improves the quality of dimension reduction data and the accuracy of subsequent processing; and 4, a dynamic topology routing mechanism enables the decision tree to be adaptively adjusted according to data characteristics, so that the classification accuracy and the processing speed are improved.

The invention discloses a brain stimulation reconstruction method based on visual experience retrieval, a medium and equipment, and relates to the technical field of computational vision, the method comprises the following steps: extracting low-level visual clues and high-level visual clues of a to-be-processed fMRI signal to ensure the retention of semantic information in a reconstruction process; by simulating the memory retrieval process of the human brain, the visual memory image is retrieved, context information is provided for degraded visual stimulation, the image reconstruction precision is further improved, and semantic details are enriched; a plurality of reconstructed images are generated by using low-level visual cues and super visual cues (a fusion result of high-level visual cues and CLIP features of visual memory images), and a final reconstructed image is selected from the reconstructed images, so that key semantic information in visual stimulation can be fully captured or reserved, and the accuracy of an image reconstruction result is improved.

The invention belongs to the technical field of fertilizer preparation, and relates to a compound fertilizer for efficiently preventing and controlling rice blast. The compound fertilizer comprises the following raw materials: a plant-source organic fertilizer, an animal-source organic fertilizer, humic acid, a functional carrier and bacillus subtilis, the plant-source organic fertilizer provides effective components such as sulfur compounds and glucosinolate, and the functional carrier can provide silicon for rice. The compound fertilizer for efficiently preventing and controlling the rice blast, prepared by the invention, not only can provide comprehensive nutrients for the growth of rice, but also can enhance the prevention and control effect on the rice blast by utilizing antibacterial and antifungal compounds contained in the plant source organic fertilizer.

The utility model relates to the technical field of agricultural machinery, in particular to a profiling type crop stalk cutting device which comprises a rack, a horizontally-arranged transmission shaft is rotatably arranged in the rack, and a plurality of vertically-arranged rotating shafts are rotatably arranged on the rack; a mounting groove is formed in the bottom end of the rotating shaft, the upper portion of a connecting shaft is arranged in the mounting groove and can move in the length direction of the rotating shaft, the upper portion of the connecting shaft is connected with the inner wall of the mounting groove through a vertically-arranged spline, a guide disc is mounted at the bottom end of the connecting shaft, and a cutting disc is fixedly connected to the connecting shaft in a sleeving mode. The straw cutting machine has the advantages that when the straw cutting machine works, under the action of gravity, the connecting shaft moves downwards to the guiding disc to be placed on the bottom face of a farmland, the height of the cutting disc is reduced along with downward movement of the connecting shaft, and therefore the cutting disc can cut off straw along the roots of the straw, complete straw can be kept, and the straw cutting machine is convenient to use. And the cut straw can be conveniently bundled, and waste of the straw is avoided.

The invention discloses a cold-chain logistics vehicle planning system. The system comprises the following steps of initial population creation, explosion and spark generation, evaluation and selection, genetic manipulation, pheromone updating, termination judgment and optimal solution output. According to the method, a dynamic pheromone updating and path reward and punishment mechanism is introduced through improvement of pheromone updating, so that the convergence speed of the algorithm is accelerated, and the quality of a solution is improved; the exploration depth of a solution space is increased by improving the crossover operation and considering the length and quality of an introduced path; improvement of mutation operation depends on a self-adaptive mutation strategy, the exploration capability of the algorithm is enhanced, multi-objective optimization is introduced into path planning through the improved firework algorithm to consider transportation cost, temperature control and energy consumption management at the same time, and a dynamic adjustment mechanism is combined to cope with frequently changing terrain and climate conditions in the karst landform, so that the development of the karst landform is facilitated. Therefore, the adaptability of the algorithm in a complex environment is improved.

The invention provides a three-dimensional reconstruction method and system in a nuclear irradiation environment, and relates to the technical field of computer vision. According to the method, a Swin Transformer network is adopted to perform feature extraction processing on an original image containing radiation noise, a constructed three-dimensional reconstruction network adopts a U-Net architecture, reservation of detail features is ensured through an encoding-decoding process, after radiation noise processing is completed, an FFB6D network is introduced to directly process a nuclear radiation environment reconstruction image, and the method has the advantages that the method is simple and convenient to operate and high in practicability. The accuracy and the stability of pose estimation under severe visual conditions are improved, and a geometric basis is provided for three-dimensional reconstruction. Model selection and optimization are carried out on all the modules adopted by the system according to the characteristics of the nuclear irradiation environment, technical resultant force is formed, a high-performance and full-automatic three-dimensional reconstruction system specially designed for the extremely severe visual environment such as nuclear irradiation is finally constructed, and the precision, robustness and practicability of three-dimensional perception in the field are remarkably improved.

The utility model discloses a sewage treatment system, which relates to the technical field of sewage treatment equipment and comprises an anaerobic biochemical reactor, an aerobic biochemical reactor and a settling tank, and both the anaerobic biochemical reactor and the aerobic biochemical reactor are connected with the settling tank; each of the anaerobic biochemical reactor and the aerobic biochemical reactor comprises a reactor main body, the reactor main body comprises a shell cylinder, and a plurality of eccentric plates which are sequentially staggered layer by layer from bottom to top are arranged in the shell cylinder; reverse-flow inclined net-shaped weir plates are arranged on the upper face and the lower face of the eccentric plate. A filler ball is also arranged on the eccentric plate; the sewage is baffled in the reactor, so that the contact time of the sewage and the sludge is prolonged, and the sewage treatment effect is improved; activated sludge films are formed on the netlike weir plates and the filler rings which are arranged on the eccentric plate at intervals, so that the area is large, and the reaction rate can be increased; net-shaped weir plates opposite to the water flow direction are arranged above and below the eccentric plate, so that the flow speed of sewage in the reactor can be reduced, and the loss of sludge biological membranes is reduced.

The invention provides a high-pressure-resistant high-power salt concentrated composite nanofiltration membrane and a preparation method thereof, and belongs to the technical field of nanofiltration membranes. The preparation method comprises the following steps: adding a terpolymer with a core-shell structure during preparation of a base membrane to prepare a membrane casting solution, scraping the membrane casting solution on a non-woven fabric to prepare the base membrane, dip-coating the prepared base membrane with an aqueous solution containing polyamine, then carrying out interfacial polymerization reaction on the base membrane, polyacyl chloride and an organic phase solution of an acyl chloride monomer with a naphthalene ring structure, and carrying out primary high-temperature cross-linking, so as to obtain the membrane. Washing with an acid solution, washing with pure water, protecting pores with glycerol and drying at high temperature to obtain the final high-pressure-resistant high-power salt concentrated composite nanofiltration membrane.

The invention belongs to the technical field of traditional Chinese medicine preparations, and particularly relates to traditional Chinese medicine granules for treating icteric hepatitis and a preparation method thereof. The traditional Chinese medicine granules are prepared from the following components: herba artemisiae scopariae, herba lysimachiae, fried fructus gardeniae, radix et rhizoma rhei, radix isatidis, herba pogostemonis, herba eupatorii, herba artemisiae annuae, radix paeoniae rubra, cortex moutan, radix salviae miltiorrhizae, fructus crataegi, semen coicis and According to the traditional Chinese medicine composition, the traditional Chinese medicines are scientifically compounded and are mutually assisted and compatible, so that the comprehensive treatment effects of clearing heat, promoting diuresis and giving consideration to exterior syndromes, stasis syndromes and distention and fullness syndromes are achieved, and the traditional Chinese medicine composition is suitable for Yang jaundice diseases in traditional Chinese medicine theories, especially for complex illness conditions with various mixed symptoms such as the exterior syndromes, the stasis syndromes and the distention and fullness syndromes. A comprehensive treatment solution is provided, the application range is wider, and the expected curative effect is more obvious.

This application relates to the field of data processing technology, and in particular to a method, apparatus, device, and storage medium for model training and 3D mesh reconstruction, used to rapidly generate high-quality 3D results. The model training method includes: obtaining an initial repair model by distilling a student model to be trained; the initial repair model is used to make the input initial sample set, after M iterative repairs, present the shape of a 3D point cloud corresponding to the initial sample set; obtaining an intermediate repair model by distilling the initial repair model; the intermediate repair model is used to make the input intermediate sample set, after N iterative repairs, present the shape of a 3D point cloud corresponding to the intermediate sample set; and obtaining a target repair model by adversarial training of the intermediate repair model based on the initial repair model, so that the error between the results obtained by repairing the same target sample set using the target repair model and the initial repair model conforms to a preset error range.

The application discloses a multifunctional injectable hydrogel and a preparation method and application thereof, and relates to the technical field of multifunctional hydrogel preparation and application. In the preparation, CMC-AGE and OD-SH are prepared respectively, and DSPE-se-se-PEG@AC NPs nanoparticles are prepared, then the multifunctional injectable hydrogel is prepared by forming a hydrogel through CMC-AGE and OD-SH and adding DSPE-se-se-PEG@AC NPs nanoparticles and / or beta nerve growth factor. The application also includes the multifunctional injectable hydrogel prepared by the above method and the application thereof. The application constructs an in-situ hydrogel delivery platform with good biocompatibility, is used for co-delivery of DSPE-se-se-PEG@AC NPs and NGF, is used in ischemic stroke treatment, and reshapes the pathological microenvironment of a cerebral infarction area, so that cerebral ischemic injury is reduced.

The application discloses a cable cooling device, which comprises a water containing cavity, the two ends of the water containing cavity are respectively provided with an inlet end and an outlet end, the water containing cavity is connected with at least one water inlet end, and the cable cooling device further comprises a cold transfer structure, a plurality of cold transfer rollers are arranged in the water containing cavity, the cold transfer rollers are hollow and internally provided with refrigerant, and the upper end of the cold transfer roller is provided with an upper liquid spraying member; and an internal circulation structure, the internal circulation structure comprises an internal circulation groove located at the bottom of the cold transfer roller, and the projection positions of the cold transfer roller and the upper liquid spraying member are located in the range of the internal circulation groove.

The embodiment of the invention relates to an information recommendation method and device, a computer program product and electronic equipment, and relates to the technical field of computers.The method comprises the steps that sample user data is obtained from a target environment, model parameters of a recommendation model are adjusted based on the sample user data, and a trained recommendation model is generated; performing zero-knowledge certification on the trained recommendation model to determine a certification result; and under the condition that the certification result is that verification is passed, performing feature fitting on user data of a target user based on the trained recommendation model, and determining recommendation information for the target user. According to the method and the device, the recommendation information for the target user can be accurately determined, and precise marketing is realized.

The present application relates to the biological field, and particularly to a weak acid buffer and its use in an immunopeptidome separation reagent, the weak acid buffer comprising amino acids and / or polypeptides, the mass-to-charge ratio of the amino acids and / or polypeptides being theoretically predictable and experimentally reproducible, or the mass-to-charge ratio being beyond the preset acquisition range of a mass spectrometer; the weak acid buffer further comprising a volatile acid, the volume concentration of the volatile acid being 0.01% to 10%; and the pH of the weak acid buffer being 3.0 to 3.5. The weak acid buffer of the present application has high compatibility, can maximally avoid pollution and signal interference to an analysis system, has a simple experimental procedure, and the overall operation can be completed within tens of minutes, thus having low technical threshold and high acceptability; can improve the number of immunopeptide groups identified, thus improving the efficiency of the experiment; and the reagent is cheap and easy to obtain, thus the method has low cost, is suitable for large-scale or even industrial use, has high universality, and the like.

The invention discloses a fault record storage device, an upper computer and a fault record query system, and the device comprises a plurality of first communication interfaces which have different communication protocols and are used for accessing target monitoring equipment so as to obtain a plurality of fault records of the target monitoring equipment based on the corresponding communication protocols; the data processing unit is connected with the plurality of first communication interfaces and is used for storing a plurality of fault records and circularly covering the plurality of fault records based on a storage sequence of the plurality of fault records under the condition that the residual storage space of the data processing unit is smaller than a preset threshold value; the second communication interface is connected with the data processing unit and used for accessing an upper computer so as to send the data messages corresponding to the multiple fault records to the upper computer for a user to inquire, effective storage of the fault records can be achieved only through a single device, then the cost of the fault records can be reduced, and the user experience is improved. And the applicability of the fault record storage device is improved.

The invention discloses an interest point recommendation method, a storage medium and equipment, and relates to the technical field of location-based social network data mining, and the method comprises the following steps: obtaining a sign-in record item sequence of a user from a location-based social network; generating item embedding vectors for the sign-in record items, constructing position embedding vectors to represent the sequence of the items, constructing frequency embedding vectors to represent the importance of repeated items, and splicing the item embedding vectors, the position embedding vectors and the frequency embedding vectors into combined embedding vectors; inputting the combined embedded vector into a bidirectional self-attention layer to obtain the representation of the combined embedded vector; according to the expressions of the item embedding vector and the combination embedding vector, using a gating vector to obtain a user intention expression; key user intention representation is obtained; on the basis of the item embedding vector, the key user intention representation is decoded and then aggregated, and interest points of the target user are generated for recommendation according to aggregated item distribution. The method provided by the invention has a remarkable effect when coping with long-sequence and multi-intention user interaction.

This invention relates to a water purification agent with post-adjusted basicity, its preparation method, and its application, belonging to the field of water purification agent technology. The invention provides a method for preparing a water purification agent with post-adjusted basicity, the specific steps of which are: adding a regulator to a water purification agent to obtain a water purification agent with post-adjusted basicity; wherein the water purification agent is an iron salt solution, specifically FeCl3 solution, and the regulator is aluminum citrate solution; or the water purification agent is an aluminum salt solution, and the regulator is at least one of an organic acid and dilute hydrochloric acid. This invention, through the method of post-adjusting basicity, changes the basicity of the water purification agent system, thereby improving the utilization rate of the water purification agent. The prepared water purification agent with post-adjusted basicity ensures water quality treatment effect and stability, and reduces the dosage, thereby reducing the cost of the reagent. In high-alkalinity water bodies, the hydrolysis rate decreases by 50%, and the reagent utilization rate increases to 88%.