CN111949708A - Multi-task prediction method, device, equipment and medium based on time sequence feature extraction - Google Patents
Multi-task prediction method, device, equipment and medium based on time sequence feature extraction Download PDFInfo
- Publication number
- CN111949708A CN111949708A CN202010797382.0A CN202010797382A CN111949708A CN 111949708 A CN111949708 A CN 111949708A CN 202010797382 A CN202010797382 A CN 202010797382A CN 111949708 A CN111949708 A CN 111949708A
- Authority
- CN
- China
- Prior art keywords
- prediction
- data
- target
- processing
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Databases & Information Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Computational Biology (AREA)
- Algebra (AREA)
- Operations Research (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Fuzzy Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the field of artificial intelligence, and provides a multi-task prediction method, a device, equipment and a medium based on time sequence feature extraction, which can identify a prediction scene from a prediction instruction, call at least one target prediction model according to the prediction scene, extract correlation features of data to be processed to obtain correlation features, continuously process the correlation features to obtain continuity features, screen and verify the continuity features to obtain target features, enable the obtained features to simultaneously follow the correlation and the continuity through multiple feature extractions, have stable performance in the model, have stronger model applicability and high usability relative to the prediction model, further input the target features into the target prediction model, output the prediction result, and combine an artificial intelligence means to enable the obtained prediction result to be more accurate and reliable. The invention also relates to a block chain technology, and the target prediction model and the prediction result can be stored in the block chain.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a multi-task prediction method, a multi-task prediction device, a multi-task prediction equipment and a multi-task prediction medium based on time sequence feature extraction.
Background
At present, many fields relate to the problem of prediction by using a model, and particularly when the problem of time series prediction is related, the correlation and continuity between feature data are not enough due to poor adopted feature data, so that data information cannot be accurately expressed, and the accuracy of model prediction is influenced.
For the above problems, algorithms such as ARIMA (differential Integrated Moving Average Autoregressive model, also called Integrated Moving Average Autoregressive model) are mainly used in the existing solution, these algorithms only predict the future value of a group of data, and in practical application, usually face multiple groups of historical service data and predict through multiple groups of different historical service data, therefore, on the premise that the traditional feature extraction method only aims at a group of data, the data information cannot be accurately expressed in practical application, and the reliability of prediction is affected.
In addition, in the prior art, only a single model is usually used for predicting the extracted features, and since the single model may be limited to data in a specific scene during training, the accuracy of prediction still needs to be improved.
Disclosure of Invention
In view of the above, it is necessary to provide a multi-task prediction method, apparatus, device and medium based on time series feature extraction, which can perform multiple feature extractions, so that the obtained features simultaneously follow correlation and continuity, and have stable performance in a model, and the model has stronger applicability, and has high availability compared with a prediction model, and further combines with an artificial intelligence means, so that the obtained prediction result is more accurate and reliable.
A multi-task prediction method based on time sequence feature extraction comprises the following steps:
identifying a prediction scene from a prediction instruction when the prediction instruction is received;
calling at least one target prediction model according to the prediction scene;
acquiring data to be processed according to the prediction instruction;
performing relevance feature extraction on the data to be processed to obtain relevance features;
carrying out continuity processing on the correlation characteristics to obtain continuity characteristics;
screening and verifying the continuity characteristics to obtain target characteristics;
and inputting the target characteristics into the at least one target prediction model for processing, and outputting a prediction result.
According to a preferred embodiment of the present invention, the identifying a prediction scenario from the prediction instruction comprises:
analyzing the method body of the prediction instruction to obtain the carried information of the prediction instruction;
acquiring a preset label, and matching the preset label in the carried information to obtain matched data;
determining the matched data as the predicted scene.
According to a preferred embodiment of the present invention, before invoking at least one target prediction model according to the prediction scenario, the multi-task prediction method based on time series feature extraction further includes:
obtaining a training sample;
splitting the training sample into a training set and a verification set;
training with the training set based on a configuration algorithm to obtain at least one initial model;
validating the at least one initial model with the validation set;
and determining the verified initial model as the at least one target prediction model, and deploying the at least one target prediction model to the block chain.
According to a preferred embodiment of the present invention, the extracting the correlation characteristics of the data to be processed to obtain the correlation characteristics includes:
determining a plurality of processing dimensions for the data to be processed, and determining the data contained in each processing dimension;
calculating the saturation of the data contained in each processing dimension;
for each processing dimension, acquiring a response proportion, an unresponsive proportion and an evidence weight of data contained in the processing dimension, calculating a difference value between the response proportion and the unresponsive proportion, and calculating a product of the difference value and the evidence weight as a first information value of the data contained in the processing dimension;
calculating noise of data contained in each processing dimension;
acquiring a first weight corresponding to saturation, a second weight corresponding to information value and a third weight corresponding to noise which are configured in advance;
calculating a weighted sum according to the saturation of the data contained in each processing dimension, the first information value of the data contained in each processing dimension, the noise of the data contained in each processing dimension, the first weight, the second weight and the third weight;
and acquiring data contained in the processing dimension with the highest weighted sum as the correlation characteristic.
According to a preferred embodiment of the present invention, the performing continuity processing on the correlation feature to obtain a continuity feature includes:
determining a plurality of processing modes for continuously processing the correlation characteristics;
calculating a second information value of the data under each processing mode;
determining the processing mode with the highest second information value as a target processing mode;
and extracting time sequence features from the correlation features in the target processing mode as the continuity features.
According to a preferred embodiment of the present invention, the screening and verifying the continuity features to obtain the target features includes:
carrying out duplicate removal processing on the continuity characteristics to obtain duplicate removal data;
determining a third information value of each of the de-duplicated data relative to the at least one target predictive model;
verifying each data according to the third information value of each data;
when the third information value of the data is greater than or equal to the configuration threshold value, determining that the data passes verification;
and integrating all verified data as the target characteristics.
According to a preferred embodiment of the present invention, the inputting the target feature into the at least one target prediction model for processing, and the outputting the prediction result includes:
splitting the target feature to obtain a first feature set and a second feature set;
inputting the first feature set into the at least one target prediction model, and outputting at least one piece of sub-prediction data;
training the at least one sub-prediction data by adopting a long-short term memory algorithm to obtain a target model;
and inputting the second feature set into the target model, and outputting the prediction result.
A temporal feature extraction based multitask prediction device, the temporal feature extraction based multitask prediction device comprising:
an identification unit configured to identify a prediction scene from a prediction instruction when the prediction instruction is received;
the calling unit is used for calling at least one target prediction model according to the prediction scene;
the acquisition unit is used for acquiring data to be processed according to the prediction instruction;
the extraction unit is used for extracting correlation characteristics of the data to be processed to obtain correlation characteristics;
the processing unit is used for carrying out continuity processing on the correlation characteristics to obtain continuity characteristics;
the processing unit is also used for screening and verifying the continuity characteristics to obtain target characteristics;
the processing unit is further configured to input the target feature into the at least one target prediction model for processing, and output a prediction result.
An electronic device, the electronic device comprising:
a memory storing at least one instruction; and
a processor executing instructions stored in the memory to implement the temporal feature extraction based multi-task prediction method.
A computer-readable storage medium having stored therein at least one instruction for execution by a processor in an electronic device to implement the method for multi-tasking prediction based on temporal feature extraction.
According to the technical scheme, the method can identify the prediction scene from the prediction instruction when receiving the prediction instruction, call at least one target prediction model according to the prediction scene, obtain the data to be processed according to the prediction instruction, extract the correlation characteristics of the data to be processed to obtain the correlation characteristics, continuously process the correlation characteristics to obtain the continuity characteristics, screen and verify the continuity characteristics to obtain the target characteristics, and after multiple times of characteristic extraction, the obtained characteristics simultaneously follow the correlation and the continuity, and have stable performance in the model, the model has stronger applicability and high availability relative to the prediction model, the target characteristics are further input into the at least one target prediction model, the prediction result is output, and artificial intelligence means is further combined, the obtained prediction result is more accurate and reliable.
Drawings
FIG. 1 is a flowchart of a multi-task prediction method based on timing feature extraction according to a preferred embodiment of the present invention.
FIG. 2 is a functional block diagram of a multi-task prediction device based on timing feature extraction according to a preferred embodiment of the present invention.
FIG. 3 is a schematic structural diagram of an electronic device implementing the multi-task prediction method based on time series feature extraction according to the preferred embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in detail with reference to the accompanying drawings and specific embodiments.
FIG. 1 is a flow chart of a multi-task prediction method based on time series feature extraction according to a preferred embodiment of the present invention. The order of the steps in the flow chart may be changed and some steps may be omitted according to different needs.
The multitask prediction method based on the time sequence feature extraction is applied to one or more electronic devices, the electronic devices are devices capable of automatically performing numerical calculation and/or information processing according to preset or stored instructions, and hardware of the electronic devices includes but is not limited to a microprocessor, an Application Specific Integrated Circuit (ASIC), a Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), an embedded device and the like.
The electronic device may be any electronic product capable of performing human-computer interaction with a user, for example, a Personal computer, a tablet computer, a smart phone, a Personal Digital Assistant (PDA), a game machine, an interactive Internet Protocol Television (IPTV), an intelligent wearable device, and the like.
The electronic device may also include a network device and/or a user device. The network device includes, but is not limited to, a single network server, a server group consisting of a plurality of network servers, or a Cloud Computing (Cloud Computing) based Cloud consisting of a large number of hosts or network servers.
The Network where the electronic device is located includes, but is not limited to, the internet, a wide area Network, a metropolitan area Network, a local area Network, a Virtual Private Network (VPN), and the like.
S10, when a prediction instruction is received, identifying a prediction scene from the prediction instruction.
Wherein the prediction instruction can be triggered by appointed personnel, such as doctors, assessment officers and the like.
The prediction scenarios may include disease diagnosis scenarios, performance assessment scenarios, and the like.
In at least one embodiment of the present invention, the identifying a prediction scenario from the prediction instruction comprises:
analyzing the method body of the prediction instruction to obtain the carried information of the prediction instruction;
acquiring a preset label, and matching the preset label in the carried information to obtain matched data;
determining the matched data as the predicted scene.
The preset tag can be configured in a user-defined mode and used for identifying the prediction scene, and the prediction scene can be accurately positioned through the preset tag.
Through the implementation mode, the prediction scene can be accurately determined according to the preset label and used as the basis of a subsequent calling model.
S11, calling at least one target prediction model according to the prediction scene.
In this embodiment, each prediction scenario has at least one corresponding prediction model, and these prediction models may be trained in advance.
Specifically, before invoking at least one target prediction model according to the prediction scenario, the multi-task prediction method based on time series feature extraction further includes:
obtaining a training sample;
splitting the training sample into a training set and a verification set;
training with the training set based on a configuration algorithm to obtain at least one initial model;
validating the at least one initial model with the validation set;
and determining the verified initial model as the at least one target prediction model, and deploying the at least one target prediction model to the block chain.
The configuration algorithm can be configured by self according to actual requirements, and the configuration algorithm includes, but is not limited to: neural network algorithm, linear regression algorithm.
In this embodiment, the at least one target prediction model is deployed in the block chain, so that the safety of the model can be effectively ensured.
Through the implementation mode, a plurality of prediction models can be obtained through training aiming at different prediction scenes respectively so as to be directly called in the subsequent use, and the data processing efficiency is improved. Meanwhile, the accuracy of data processing can be further improved by the combined action of the plurality of prediction models.
And S12, acquiring the data to be processed according to the prediction instruction.
In this embodiment, the obtaining of the data to be processed according to the prediction instruction includes, but is not limited to, one or a combination of the following manners:
(1) and identifying keywords from the prediction instruction, and crawling the data to be processed from a specified website through a web crawler technology according to the keywords.
Wherein the designated websites may include all websites associated with the predicted scenario.
For example: when the prediction scenario is a disease diagnosis scenario, the designated website may be various websites of a hospital.
(2) And searching data corresponding to the keywords in a specified database to serve as the data to be processed.
Wherein the specified database may include, but is not limited to: a time series feature library, a feature library associated with the predicted scene.
S13, performing correlation feature extraction on the data to be processed to obtain correlation features.
It will be appreciated that for some common prediction problems, the data employed for prediction is generally relevant and follows the principle of continuity.
In this embodiment, the correlation means that the feature and the prediction result have a correlation at the current time.
In at least one embodiment of the present invention, the performing correlation feature extraction on the data to be processed to obtain a correlation feature includes:
determining a plurality of processing dimensions for the data to be processed, and determining the data contained in each processing dimension;
calculating the saturation of the data contained in each processing dimension;
for each processing dimension, acquiring a response proportion, an unresponsive proportion and an evidence weight of data contained in the processing dimension, calculating a difference Value between the response proportion and the unresponsive proportion, and calculating a product of the difference Value and the evidence weight as a first Information Value (IV Value) of the data contained in the processing dimension;
calculating noise of data contained in each processing dimension;
acquiring a first weight corresponding to saturation, a second weight corresponding to information value and a third weight corresponding to noise which are configured in advance;
calculating a weighted sum according to the saturation of the data contained in each processing dimension, the first information value of the data contained in each processing dimension, the noise of the data contained in each processing dimension, the first weight, the second weight and the third weight;
and acquiring data contained in the processing dimension with the highest weighted sum as the correlation characteristic.
Wherein the processing dimension may be divided on a time basis. For example: the data for the 7 th month of 2018 may be set to one process dimension.
The saturation refers to the proportion of valid data after the null value is eliminated.
The noise may include an influence of an invalid operation such as an invalid click.
The first weight, the second weight, and the third weight may be obtained through experiments, that is, an optimal weight configuration manner is determined through experiments and configured according to the optimal weight configuration manner.
In this embodiment, when the correlation feature is extracted, the targeted variable is an independent variable at a certain time and a dependent variable at a certain time. For example: the independent variable is the class rate of 7 months, the examination result of 7 months, and the dependent variable is the examination result of 7 months at the end of the period.
Through the implementation mode, the first-layer screening of the features can be realized, the correlation features with high correlation with the prediction result are extracted from the data to be processed, the generation of irrelevant noise and unsaturated features is effectively restrained, the correlation dimensions are extremely fast enriched, the extracted features have good correlation, and the accuracy of the prediction result is further improved.
And S14, carrying out continuity processing on the correlation characteristics to obtain continuity characteristics.
In the present embodiment, continuity means that the extracted features follow the principle of continuity and can be used to predict future events.
In at least one embodiment of the present invention, the performing continuity processing on the correlation feature to obtain a continuity feature includes:
determining a plurality of processing modes for continuously processing the correlation characteristics;
calculating a second information value of the data under each processing mode;
determining the processing mode with the highest second information value as a target processing mode;
and extracting time sequence features from the correlation features in the target processing mode as the continuity features.
For example: and when the second information value is determined to be the highest by calculation when the feature extraction is carried out by using a tsfresh tool, extracting a time sequence feature from the correlation feature by using the tsfresh tool as the continuity feature.
In this embodiment, when the continuity feature is extracted, the variable to be targeted is an independent variable of a certain period and a dependent variable of the next time corresponding to the certain period. For example: the independent variable is the class rate of 7-8 months, the examination result of 7-8 months, and the dependent variable is the examination result of 9 months at the end of the period.
Through the embodiment, the second-layer screening of the features can be realized, and the continuity features are further extracted from the correlation features, so that the extracted features are ensured to follow the continuity principle.
And S15, screening and verifying the continuity characteristics to obtain target characteristics.
It should be noted that, through the above-mentioned correlation feature extraction and continuity feature extraction, there may be unavailability of the obtained features.
For example: the extracted features may have high repeatability or have low adaptability to the prediction model, which may affect the performance of the features in model prediction, and in addition, whether the performance of data in the model is stable also affects the final prediction result and prediction efficiency, so the embodiment also needs to screen and verify the extracted continuous features to extract the features that can have stable prediction performance in the model.
Specifically, the screening and verifying the continuity features to obtain the target features includes:
carrying out duplicate removal processing on the continuity characteristics to obtain duplicate removal data;
determining a third information value of each of the de-duplicated data relative to the at least one target predictive model;
verifying each data according to the third information value of each data;
when the third information value of the data is greater than or equal to the configuration threshold value, determining that the data passes verification;
and integrating all verified data as the target characteristics.
The configuration threshold value can be configured in a user-defined mode according to actual requirements.
In this embodiment, when the target feature is extracted, the targeted variable is an independent variable of a certain period and a dependent variable of a next period corresponding to the certain period. For example: the independent variable is the class rate of 7-8 months, the examination result of 7-8 months, and the dependent variable is the examination result of 9-10 months.
Through the implementation mode, the third-layer screening of the features can be realized, and the target features which are stably expressed and available in the prediction model, namely the features with stronger model applicability, are further extracted from the continuity features, so that the modeling capability is improved, and the reliability of the prediction result is further ensured.
In addition, the three-time feature extraction process in the embodiment can be executed independently, so that each link can be optimized in stages, and the method has more logicality and interpretability so as to find data problems in time.
And S16, inputting the target characteristics into the at least one target prediction model for processing, and outputting a prediction result.
In the embodiment, the target features simultaneously follow the correlation and the continuity, have stable performance in the model and have high availability relative to the prediction model, so that the prediction process can be more efficient and the prediction result is more accurate.
In order to ensure that data is not tampered with, the prediction result may be stored in a blockchain, so as to improve the security and privacy of data.
In at least one embodiment of the present invention, the inputting the target feature into the at least one target prediction model for processing, and the outputting the prediction result includes:
splitting the target feature to obtain a first feature set and a second feature set;
inputting the first feature set into the at least one target prediction model, and outputting at least one piece of sub-prediction data;
training the at least one sub-prediction data by adopting a long-short term memory algorithm to obtain a target model;
and inputting the second feature set into the target model, and outputting the prediction result.
Wherein, the Long Short Term Memory algorithm (LSTM) comprises three network layers, which are respectively: an input gate layer, a forgetting gate layer and an output gate layer.
Through the implementation mode, the long-short term memory algorithm has the advantage of time series, so that the target model trained by the long-short term memory algorithm also has certain time sequence, and the time sequence characteristics can be better processed.
In addition, the target characteristics are split, two-stage prediction is performed based on the split characteristics, and the prediction accuracy is further improved.
In at least one embodiment of the present invention, the training the at least one sub-prediction data using a long-short term memory algorithm to obtain the target model comprises:
inputting the at least one sub-prediction data into the forgetting gate layer to carry out forgetting processing to obtain first data;
dividing the first data into a second data set and a third data set by adopting a cross verification method;
inputting the second data set to the input gate layer for training to obtain a secondary learner;
and verifying the secondary learner by the third data set to obtain the target model.
Through the implementation mode, the target model can be obtained based on the long-term and short-term memory algorithm training for subsequent prediction.
Specifically, dividing the first data into the second data set and the third data set by using a cross-validation method includes:
randomly dividing the first data into at least one data packet according to a preset number, determining any one data packet in the at least one data packet as the third data set, determining the rest data packets as the second data set, and repeating the above processes until all the data packets are sequentially used as the third data set.
Through the embodiment, full-scale training and verification can be performed by using all the first data, and the fitting degree of model training is improved.
According to the technical scheme, the method can identify the prediction scene from the prediction instruction when receiving the prediction instruction, call at least one target prediction model according to the prediction scene, obtain the data to be processed according to the prediction instruction, extract the correlation characteristics of the data to be processed to obtain the correlation characteristics, continuously process the correlation characteristics to obtain the continuity characteristics, screen and verify the continuity characteristics to obtain the target characteristics, and after multiple times of characteristic extraction, the obtained characteristics simultaneously follow the correlation and the continuity, and have stable performance in the model, the model has stronger applicability and high availability relative to the prediction model, the target characteristics are further input into the at least one target prediction model, the prediction result is output, and artificial intelligence means is further combined, the obtained prediction result is more accurate and reliable.
FIG. 2 is a functional block diagram of a multi-task prediction apparatus based on timing feature extraction according to a preferred embodiment of the present invention. The multitask prediction device 11 based on the time-series feature extraction comprises a recognition unit 110, a calling unit 111, an acquisition unit 112, an extraction unit 113 and a processing unit 114. The module/unit referred to in the present invention refers to a series of computer program segments that can be executed by the processor 13 and that can perform a fixed function, and that are stored in the memory 12. In the present embodiment, the functions of the modules/units will be described in detail in the following embodiments.
When a prediction instruction is received, the recognition unit 110 recognizes a prediction scene from the prediction instruction.
Wherein the prediction instruction can be triggered by appointed personnel, such as doctors, assessment officers and the like.
The prediction scenarios may include disease diagnosis scenarios, performance assessment scenarios, and the like.
In at least one embodiment of the present invention, the identifying unit 110 identifies a prediction scenario from the prediction instruction includes:
analyzing the method body of the prediction instruction to obtain the carried information of the prediction instruction;
acquiring a preset label, and matching the preset label in the carried information to obtain matched data;
determining the matched data as the predicted scene.
The preset tag can be configured in a user-defined mode and used for identifying the prediction scene, and the prediction scene can be accurately positioned through the preset tag.
Through the implementation mode, the prediction scene can be accurately determined according to the preset label and used as the basis of a subsequent calling model.
The retrieval unit 111 retrieves at least one target prediction model from the prediction scenario.
In this embodiment, each prediction scenario has at least one corresponding prediction model, and these prediction models may be trained in advance.
Specifically, before at least one target prediction model is called according to the prediction scene, a training sample is obtained;
splitting the training sample into a training set and a verification set;
training with the training set based on a configuration algorithm to obtain at least one initial model;
validating the at least one initial model with the validation set;
and determining the verified initial model as the at least one target prediction model, and deploying the at least one target prediction model to the block chain.
The configuration algorithm can be configured by self according to actual requirements, and the configuration algorithm includes, but is not limited to: neural network algorithm, linear regression algorithm.
In this embodiment, the at least one target prediction model is deployed in the block chain, so that the safety of the model can be effectively ensured.
Through the implementation mode, a plurality of prediction models can be obtained through training aiming at different prediction scenes respectively so as to be directly called in the subsequent use, and the data processing efficiency is improved. Meanwhile, the accuracy of data processing can be further improved by the combined action of the plurality of prediction models.
The fetch unit 112 fetches the data to be processed according to the prediction instruction.
In this embodiment, the obtaining unit 112 obtains the data to be processed according to the prediction instruction, which includes, but is not limited to, one or a combination of the following ways:
(1) and identifying keywords from the prediction instruction, and crawling the data to be processed from a specified website through a web crawler technology according to the keywords.
Wherein the designated websites may include all websites associated with the predicted scenario.
For example: when the prediction scenario is a disease diagnosis scenario, the designated website may be various websites of a hospital.
(2) And searching data corresponding to the keywords in a specified database to serve as the data to be processed.
Wherein the specified database may include, but is not limited to: a time series feature library, a feature library associated with the predicted scene.
The extracting unit 113 performs correlation feature extraction on the data to be processed to obtain correlation features.
It will be appreciated that for some common prediction problems, the data employed for prediction is generally relevant and follows the principle of continuity.
In this embodiment, the correlation means that the feature and the prediction result have a correlation at the current time.
In at least one embodiment of the present invention, the extracting unit 113 performs correlation feature extraction on the data to be processed, and obtaining a correlation feature includes:
determining a plurality of processing dimensions for the data to be processed, and determining the data contained in each processing dimension;
calculating the saturation of the data contained in each processing dimension;
for each processing dimension, acquiring a response proportion, an unresponsive proportion and an evidence weight of data contained in the processing dimension, calculating a difference Value between the response proportion and the unresponsive proportion, and calculating a product of the difference Value and the evidence weight as a first Information Value (IV Value) of the data contained in the processing dimension;
calculating noise of data contained in each processing dimension;
acquiring a first weight corresponding to saturation, a second weight corresponding to information value and a third weight corresponding to noise which are configured in advance;
calculating a weighted sum according to the saturation of the data contained in each processing dimension, the first information value of the data contained in each processing dimension, the noise of the data contained in each processing dimension, the first weight, the second weight and the third weight;
and acquiring data contained in the processing dimension with the highest weighted sum as the correlation characteristic.
Wherein the processing dimension may be divided on a time basis. For example: the data for the 7 th month of 2018 may be set to one process dimension.
The saturation refers to the proportion of valid data after the null value is eliminated.
The noise may include an influence of an invalid operation such as an invalid click.
The first weight, the second weight, and the third weight may be obtained through experiments, that is, an optimal weight configuration manner is determined through experiments and configured according to the optimal weight configuration manner.
In this embodiment, when the correlation feature is extracted, the targeted variable is an independent variable at a certain time and a dependent variable at a certain time. For example: the independent variable is the class rate of 7 months, the examination result of 7 months, and the dependent variable is the examination result of 7 months at the end of the period.
Through the implementation mode, the first-layer screening of the features can be realized, the correlation features with high correlation with the prediction result are extracted from the data to be processed, the generation of irrelevant noise and unsaturated features is effectively restrained, the correlation dimensions are extremely fast enriched, the extracted features have good correlation, and the accuracy of the prediction result is further improved.
The processing unit 114 performs continuity processing on the correlation characteristic to obtain a continuity characteristic.
In the present embodiment, continuity means that the extracted features follow the principle of continuity and can be used to predict future events.
In at least one embodiment of the present invention, the processing unit 114 performs continuity processing on the correlation feature, and obtaining the continuity feature includes:
determining a plurality of processing modes for continuously processing the correlation characteristics;
calculating a second information value of the data under each processing mode;
determining the processing mode with the highest second information value as a target processing mode;
and extracting time sequence features from the correlation features in the target processing mode as the continuity features.
For example: and when the second information value is determined to be the highest by calculation when the feature extraction is carried out by using a tsfresh tool, extracting a time sequence feature from the correlation feature by using the tsfresh tool as the continuity feature.
In this embodiment, when the continuity feature is extracted, the variable to be targeted is an independent variable of a certain period and a dependent variable of the next time corresponding to the certain period. For example: the independent variable is the class rate of 7-8 months, the examination result of 7-8 months, and the dependent variable is the examination result of 9 months at the end of the period.
Through the embodiment, the second-layer screening of the features can be realized, and the continuity features are further extracted from the correlation features, so that the extracted features are ensured to follow the continuity principle.
The processing unit 114 performs screening and verification processing on the continuity features to obtain target features.
It should be noted that, through the above-mentioned correlation feature extraction and continuity feature extraction, there may be unavailability of the obtained features.
For example: the extracted features may have high repeatability or have low adaptability to the prediction model, which may affect the performance of the features in model prediction, and in addition, whether the performance of data in the model is stable also affects the final prediction result and prediction efficiency, so the embodiment also needs to screen and verify the extracted continuous features to extract the features that can have stable prediction performance in the model.
Specifically, the processing unit 114 performs screening and verification processing on the continuity features, and obtaining target features includes:
carrying out duplicate removal processing on the continuity characteristics to obtain duplicate removal data;
determining a third information value of each of the de-duplicated data relative to the at least one target predictive model;
verifying each data according to the third information value of each data;
when the third information value of the data is greater than or equal to the configuration threshold value, determining that the data passes verification;
and integrating all verified data as the target characteristics.
The configuration threshold value can be configured in a user-defined mode according to actual requirements.
In this embodiment, when the target feature is extracted, the targeted variable is an independent variable of a certain period and a dependent variable of a next period corresponding to the certain period. For example: the independent variable is the class rate of 7-8 months, the examination result of 7-8 months, and the dependent variable is the examination result of 9-10 months.
Through the implementation mode, the third-layer screening of the features can be realized, and the target features which are stably expressed and available in the prediction model, namely the features with stronger model applicability, are further extracted from the continuity features, so that the modeling capability is improved, and the reliability of the prediction result is further ensured.
In addition, the three-time feature extraction process in the embodiment can be executed independently, so that each link can be optimized in stages, and the method has more logicality and interpretability so as to find data problems in time.
The processing unit 114 inputs the target feature into the at least one target prediction model for processing, and outputs a prediction result.
In the embodiment, the target features simultaneously follow the correlation and the continuity, have stable performance in the model and have high availability relative to the prediction model, so that the prediction process can be more efficient and the prediction result is more accurate.
In order to ensure that data is not tampered with, the prediction result may be stored in a blockchain, so as to improve the security and privacy of data.
In at least one embodiment of the present invention, the processing unit 114 inputs the target feature into the at least one target prediction model for processing, and outputting the prediction result includes:
splitting the target feature to obtain a first feature set and a second feature set;
inputting the first feature set into the at least one target prediction model, and outputting at least one piece of sub-prediction data;
training the at least one sub-prediction data by adopting a long-short term memory algorithm to obtain a target model;
and inputting the second feature set into the target model, and outputting the prediction result.
Wherein, the Long Short Term Memory algorithm (LSTM) comprises three network layers, which are respectively: an input gate layer, a forgetting gate layer and an output gate layer.
Through the implementation mode, the long-short term memory algorithm has the advantage of time series, so that the target model trained by the long-short term memory algorithm also has certain time sequence, and the time sequence characteristics can be better processed.
In addition, the target characteristics are split, two-stage prediction is performed based on the split characteristics, and the prediction accuracy is further improved.
In at least one embodiment of the present invention, the training the at least one sub-prediction data using a long-short term memory algorithm to obtain the target model comprises:
inputting the at least one sub-prediction data into the forgetting gate layer to carry out forgetting processing to obtain first data;
dividing the first data into a second data set and a third data set by adopting a cross verification method;
inputting the second data set to the input gate layer for training to obtain a secondary learner;
and verifying the secondary learner by the third data set to obtain the target model.
Through the implementation mode, the target model can be obtained based on the long-term and short-term memory algorithm training for subsequent prediction.
Specifically, dividing the first data into the second data set and the third data set by using a cross-validation method includes:
randomly dividing the first data into at least one data packet according to a preset number, determining any one data packet in the at least one data packet as the third data set, determining the rest data packets as the second data set, and repeating the above processes until all the data packets are sequentially used as the third data set.
Through the embodiment, full-scale training and verification can be performed by using all the first data, and the fitting degree of model training is improved.
According to the technical scheme, the method can identify the prediction scene from the prediction instruction when receiving the prediction instruction, call at least one target prediction model according to the prediction scene, obtain the data to be processed according to the prediction instruction, extract the correlation characteristics of the data to be processed to obtain the correlation characteristics, continuously process the correlation characteristics to obtain the continuity characteristics, screen and verify the continuity characteristics to obtain the target characteristics, and after multiple times of characteristic extraction, the obtained characteristics simultaneously follow the correlation and the continuity, and have stable performance in the model, the model has stronger applicability and high availability relative to the prediction model, the target characteristics are further input into the at least one target prediction model, the prediction result is output, and artificial intelligence means is further combined, the obtained prediction result is more accurate and reliable.
Fig. 3 is a schematic structural diagram of an electronic device implementing the multi-task prediction method based on time series feature extraction according to the preferred embodiment of the present invention.
The electronic device 1 may comprise a memory 12, a processor 13 and a bus, and may further comprise a computer program, such as a multi-task prediction program based on timing feature extraction, stored in the memory 12 and executable on the processor 13.
It will be understood by those skilled in the art that the schematic diagram is merely an example of the electronic device 1, and does not constitute a limitation to the electronic device 1, the electronic device 1 may have a bus-type structure or a star-type structure, the electronic device 1 may further include more or less hardware or software than those shown in the figures, or different component arrangements, for example, the electronic device 1 may further include an input and output device, a network access device, and the like.
It should be noted that the electronic device 1 is only an example, and other existing or future electronic products, such as those that can be adapted to the present invention, should also be included in the scope of the present invention, and are included herein by reference.
The memory 12 includes at least one type of readable storage medium, which includes flash memory, removable hard disks, multimedia cards, card-type memory (e.g., SD or DX memory, etc.), magnetic memory, magnetic disks, optical disks, etc. The memory 12 may in some embodiments be an internal storage unit of the electronic device 1, for example a removable hard disk of the electronic device 1. The memory 12 may also be an external storage device of the electronic device 1 in other embodiments, such as a plug-in mobile hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided on the electronic device 1. Further, the memory 12 may also include both an internal storage unit and an external storage device of the electronic device 1. The memory 12 may be used not only to store application software installed in the electronic device 1 and various types of data, such as codes of a multitask prediction program based on time-series feature extraction, etc., but also to temporarily store data that has been output or is to be output.
The processor 13 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same or different functions, including one or more Central Processing Units (CPUs), microprocessors, digital Processing chips, graphics processors, and combinations of various control chips. The processor 13 is a Control Unit (Control Unit) of the electronic device 1, connects various components of the electronic device 1 by using various interfaces and lines, and executes various functions and processes data of the electronic device 1 by running or executing programs or modules (for example, executing a multitask prediction program based on time-series feature extraction, and the like) stored in the memory 12 and calling data stored in the memory 12.
The processor 13 executes an operating system of the electronic device 1 and various installed application programs. The processor 13 executes the application program to implement the steps in each of the above embodiments of the multi-task prediction method based on time series feature extraction, such as the steps shown in fig. 1.
Illustratively, the computer program may be divided into one or more modules/units, which are stored in the memory 12 and executed by the processor 13 to accomplish the present invention. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution process of the computer program in the electronic device 1. For example, the computer program may be divided into a recognition unit 110, a call unit 111, an acquisition unit 112, an extraction unit 113, a processing unit 114.
The integrated unit implemented in the form of a software functional module may be stored in a computer-readable storage medium. The software functional module is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a computer device, or a network device) or a processor (processor) to execute parts of the multi-task prediction method based on the time series feature extraction according to the embodiments of the present invention.
The integrated modules/units of the electronic device 1 may be stored in a computer-readable storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, all or part of the flow of the method according to the embodiments of the present invention may be implemented by a computer program, which may be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method embodiments may be implemented.
Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying said computer program code, recording medium, U-disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM).
Further, the computer-usable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the blockchain node, and the like.
The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
The bus may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one arrow is shown in FIG. 3, but this does not indicate only one bus or one type of bus. The bus is arranged to enable connection communication between the memory 12 and at least one processor 13 or the like.
Although not shown, the electronic device 1 may further include a power supply (such as a battery) for supplying power to each component, and preferably, the power supply may be logically connected to the at least one processor 13 through a power management device, so as to implement functions of charge management, discharge management, power consumption management, and the like through the power management device. The power supply may also include any component of one or more dc or ac power sources, recharging devices, power failure detection circuitry, power converters or inverters, power status indicators, and the like. The electronic device 1 may further include various sensors, a bluetooth module, a Wi-Fi module, and the like, which are not described herein again.
Further, the electronic device 1 may further include a network interface, and optionally, the network interface may include a wired interface and/or a wireless interface (such as a WI-FI interface, a bluetooth interface, etc.), which are generally used for establishing a communication connection between the electronic device 1 and other electronic devices.
Optionally, the electronic device 1 may further comprise a user interface, which may be a Display (Display), an input unit (such as a Keyboard), and optionally a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, or the like. The display, which may also be referred to as a display screen or display unit, is suitable for displaying information processed in the electronic device 1 and for displaying a visualized user interface, among other things.
It is to be understood that the described embodiments are for purposes of illustration only and that the scope of the appended claims is not limited to such structures.
Fig. 3 only shows the electronic device 1 with components 12-13, and it will be understood by a person skilled in the art that the structure shown in fig. 3 does not constitute a limitation of the electronic device 1, and may comprise fewer or more components than shown, or a combination of certain components, or a different arrangement of components.
With reference to fig. 1, the memory 12 of the electronic device 1 stores a plurality of instructions to implement a multi-task prediction method based on time series feature extraction, and the processor 13 can execute the plurality of instructions to implement:
identifying a prediction scene from a prediction instruction when the prediction instruction is received;
calling at least one target prediction model according to the prediction scene;
acquiring data to be processed according to the prediction instruction;
performing relevance feature extraction on the data to be processed to obtain relevance features;
carrying out continuity processing on the correlation characteristics to obtain continuity characteristics;
screening and verifying the continuity characteristics to obtain target characteristics;
and inputting the target characteristics into the at least one target prediction model for processing, and outputting a prediction result.
Specifically, the processor 13 may refer to the description of the relevant steps in the embodiment corresponding to fig. 1 for a specific implementation method of the instruction, which is not described herein again.
In the embodiments provided in the present invention, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof.
The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or means recited in the system claims may also be implemented by one unit or means in software or hardware. The terms second, etc. are used to denote names, but not any particular order.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (10)
1. A multi-task prediction method based on time sequence feature extraction is characterized by comprising the following steps:
identifying a prediction scene from a prediction instruction when the prediction instruction is received;
calling at least one target prediction model according to the prediction scene;
acquiring data to be processed according to the prediction instruction;
performing relevance feature extraction on the data to be processed to obtain relevance features;
carrying out continuity processing on the correlation characteristics to obtain continuity characteristics;
screening and verifying the continuity characteristics to obtain target characteristics;
and inputting the target characteristics into the at least one target prediction model for processing, and outputting a prediction result.
2. The method of multi-tasking based on temporal feature extraction as recited in claim 1, wherein said identifying a prediction scenario from said prediction instruction comprises:
analyzing the method body of the prediction instruction to obtain the carried information of the prediction instruction;
acquiring a preset label, and matching the preset label in the carried information to obtain matched data;
determining the matched data as the predicted scene.
3. The time series feature extraction based multitask prediction method according to claim 1, wherein before invoking at least one target prediction model according to said prediction scenario, said time series feature extraction based multitask prediction method further comprises:
obtaining a training sample;
splitting the training sample into a training set and a verification set;
training with the training set based on a configuration algorithm to obtain at least one initial model;
validating the at least one initial model with the validation set;
and determining the verified initial model as the at least one target prediction model, and deploying the at least one target prediction model to the block chain.
4. The multi-task prediction method based on time series feature extraction as claimed in claim 1, wherein the performing correlation feature extraction on the data to be processed to obtain correlation features comprises:
determining a plurality of processing dimensions for the data to be processed, and determining the data contained in each processing dimension;
calculating the saturation of the data contained in each processing dimension;
for each processing dimension, acquiring a response proportion, an unresponsive proportion and an evidence weight of data contained in the processing dimension, calculating a difference value between the response proportion and the unresponsive proportion, and calculating a product of the difference value and the evidence weight as a first information value of the data contained in the processing dimension;
calculating noise of data contained in each processing dimension;
acquiring a first weight corresponding to saturation, a second weight corresponding to information value and a third weight corresponding to noise which are configured in advance;
calculating a weighted sum according to the saturation of the data contained in each processing dimension, the first information value of the data contained in each processing dimension, the noise of the data contained in each processing dimension, the first weight, the second weight and the third weight;
and acquiring data contained in the processing dimension with the highest weighted sum as the correlation characteristic.
5. The multi-task prediction method based on time series feature extraction as claimed in claim 1, wherein the performing continuity processing on the correlation feature to obtain a continuity feature comprises:
determining a plurality of processing modes for continuously processing the correlation characteristics;
calculating a second information value of the data under each processing mode;
determining the processing mode with the highest second information value as a target processing mode;
and extracting time sequence features from the correlation features in the target processing mode as the continuity features.
6. The multi-task prediction method based on time series feature extraction as claimed in claim 1, wherein the step of screening and verifying the continuity features to obtain the target features comprises:
carrying out duplicate removal processing on the continuity characteristics to obtain duplicate removal data;
determining a third information value of each of the de-duplicated data relative to the at least one target predictive model;
verifying each data according to the third information value of each data;
when the third information value of the data is greater than or equal to the configuration threshold value, determining that the data passes verification;
and integrating all verified data as the target characteristics.
7. The multi-task prediction method based on time series feature extraction as claimed in claim 1, wherein the inputting the target feature into the at least one target prediction model for processing, and the outputting the prediction result comprises:
splitting the target feature to obtain a first feature set and a second feature set;
inputting the first feature set into the at least one target prediction model, and outputting at least one piece of sub-prediction data;
training the at least one sub-prediction data by adopting a long-short term memory algorithm to obtain a target model;
and inputting the second feature set into the target model, and outputting the prediction result.
8. A multi-task prediction device based on time series feature extraction is characterized by comprising:
an identification unit configured to identify a prediction scene from a prediction instruction when the prediction instruction is received;
the calling unit is used for calling at least one target prediction model according to the prediction scene;
the acquisition unit is used for acquiring data to be processed according to the prediction instruction;
the extraction unit is used for extracting correlation characteristics of the data to be processed to obtain correlation characteristics;
the processing unit is used for carrying out continuity processing on the correlation characteristics to obtain continuity characteristics;
the processing unit is also used for screening and verifying the continuity characteristics to obtain target characteristics;
the processing unit is further configured to input the target feature into the at least one target prediction model for processing, and output a prediction result.
9. An electronic device, characterized in that the electronic device comprises:
a memory storing at least one instruction; and
a processor executing instructions stored in the memory to implement the method of multi-tasking prediction based on temporal feature extraction of any of claims 1-7.
10. A computer-readable storage medium characterized by: the computer-readable storage medium stores at least one instruction, which is executed by a processor in an electronic device to implement the multi-task prediction method based on time-series feature extraction according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010797382.0A CN111949708B (en) | 2020-08-10 | 2020-08-10 | Multi-task prediction method, device, equipment and medium based on time sequence feature extraction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010797382.0A CN111949708B (en) | 2020-08-10 | 2020-08-10 | Multi-task prediction method, device, equipment and medium based on time sequence feature extraction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111949708A true CN111949708A (en) | 2020-11-17 |
| CN111949708B CN111949708B (en) | 2023-07-25 |
Family
ID=73333191
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010797382.0A Active CN111949708B (en) | 2020-08-10 | 2020-08-10 | Multi-task prediction method, device, equipment and medium based on time sequence feature extraction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111949708B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112330080A (en) * | 2021-01-07 | 2021-02-05 | 平安科技（深圳）有限公司 | Factor screening method, device, equipment and medium based on connectivity graph |
| CN113076850A (en) * | 2021-03-29 | 2021-07-06 | Oppo广东移动通信有限公司 | Multitask prediction method, multitask prediction device and electronic equipment |
| CN114841267A (en) * | 2022-05-05 | 2022-08-02 | 北京百度网讯科技有限公司 | Real-time prediction method and device, electronic equipment and computer program product |
| CN115293082A (en) * | 2022-09-30 | 2022-11-04 | 深圳鸿芯微纳技术有限公司 | Training and predicting method, device, equipment and storage medium of time sequence prediction model |
| CN117373584A (en) * | 2023-12-08 | 2024-01-09 | 北京大学第一医院 | Dynamic prediction method and system for acute kidney injury |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107515898A (en) * | 2017-07-22 | 2017-12-26 | 复旦大学 | Based on data diversity and the multifarious tire enterprise sales forecasting method of task |
| US20180060665A1 (en) * | 2016-08-29 | 2018-03-01 | Nec Laboratories America, Inc. | Dual Stage Attention Based Recurrent Neural Network for Time Series Prediction |
| CN110619423A (en) * | 2019-08-06 | 2019-12-27 | 平安科技（深圳）有限公司 | Multitask prediction method and device, electronic equipment and storage medium |
-
2020
- 2020-08-10 CN CN202010797382.0A patent/CN111949708B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180060665A1 (en) * | 2016-08-29 | 2018-03-01 | Nec Laboratories America, Inc. | Dual Stage Attention Based Recurrent Neural Network for Time Series Prediction |
| CN107515898A (en) * | 2017-07-22 | 2017-12-26 | 复旦大学 | Based on data diversity and the multifarious tire enterprise sales forecasting method of task |
| CN110619423A (en) * | 2019-08-06 | 2019-12-27 | 平安科技（深圳）有限公司 | Multitask prediction method and device, electronic equipment and storage medium |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112330080A (en) * | 2021-01-07 | 2021-02-05 | 平安科技（深圳）有限公司 | Factor screening method, device, equipment and medium based on connectivity graph |
| CN113076850A (en) * | 2021-03-29 | 2021-07-06 | Oppo广东移动通信有限公司 | Multitask prediction method, multitask prediction device and electronic equipment |
| CN114841267A (en) * | 2022-05-05 | 2022-08-02 | 北京百度网讯科技有限公司 | Real-time prediction method and device, electronic equipment and computer program product |
| CN114841267B (en) * | 2022-05-05 | 2023-09-01 | 北京百度网讯科技有限公司 | Real-time prediction method, device, electronic equipment and computer program product |
| CN115293082A (en) * | 2022-09-30 | 2022-11-04 | 深圳鸿芯微纳技术有限公司 | Training and predicting method, device, equipment and storage medium of time sequence prediction model |
| CN115293082B (en) * | 2022-09-30 | 2023-01-20 | 深圳鸿芯微纳技术有限公司 | Training and predicting method, device, equipment and storage medium of time sequence prediction model |
| CN117373584A (en) * | 2023-12-08 | 2024-01-09 | 北京大学第一医院 | Dynamic prediction method and system for acute kidney injury |
| CN117373584B (en) * | 2023-12-08 | 2024-03-12 | 北京大学第一医院 | Dynamic prediction method and system for acute kidney injury |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111949708B (en) | 2023-07-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111949708B (en) | Multi-task prediction method, device, equipment and medium based on time sequence feature extraction | |
| CN112446025A (en) | Federal learning defense method and device, electronic equipment and storage medium | |
| CN112231586A (en) | Course recommendation method, device, equipment and medium based on transfer learning | |
| CN112447189A (en) | Voice event detection method and device, electronic equipment and computer storage medium | |
| CN112559535B (en) | Multithreading-based asynchronous task processing method, device, equipment and medium | |
| CN111950621A (en) | Target data detection method, device, equipment and medium based on artificial intelligence | |
| CN113806434B (en) | Big data processing method, device, equipment and medium | |
| CN111985545B (en) | Target data detection method, device, equipment and medium based on artificial intelligence | |
| CN111950707B (en) | Behavior prediction method, device, equipment and medium based on behavior co-occurrence network | |
| CN114185776A (en) | Big data point burying method, device, equipment and medium for application program | |
| CN112396547B (en) | Course recommendation method, device, equipment and medium based on unsupervised learning | |
| CN114022841A (en) | Personnel monitoring and identifying method and device, electronic equipment and readable storage medium | |
| CN113256181A (en) | Risk factor prediction method, device, equipment and medium | |
| CN112948275A (en) | Test data generation method, device, equipment and storage medium | |
| CN112101191A (en) | Expression recognition method, device, equipment and medium based on frame attention network | |
| CN114816371B (en) | Message processing method, device, equipment and medium | |
| CN116823437A (en) | Access method, device, equipment and medium based on configured wind control strategy | |
| CN112330080B (en) | Factor screening method, device, equipment and medium based on connectivity graph | |
| CN111859985B (en) | AI customer service model test method and device, electronic equipment and storage medium | |
| CN113204962A (en) | Word sense disambiguation method, device, equipment and medium based on graph expansion structure | |
| CN114996386A (en) | Business role identification method, device, equipment and storage medium | |
| CN114239538A (en) | Assertion processing method and device, computer equipment and storage medium | |
| CN114881313A (en) | Behavior prediction method and device based on artificial intelligence and related equipment | |
| CN114372892A (en) | Payment data monitoring method, device, equipment and medium | |
| CN113806540A (en) | Text labeling method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |