CN118152430A - Large language model-based farm data index analysis method, device and equipment - Google Patents
Large language model-based farm data index analysis method, device and equipment Download PDFInfo
- Publication number
- CN118152430A CN118152430A CN202410586352.3A CN202410586352A CN118152430A CN 118152430 A CN118152430 A CN 118152430A CN 202410586352 A CN202410586352 A CN 202410586352A CN 118152430 A CN118152430 A CN 118152430A
- Authority
- CN
- China
- Prior art keywords
- data
- analysis
- language model
- farm
- api
- 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.)
- Pending
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 104
- 238000000034 method Methods 0.000 claims abstract description 47
- 238000012545 processing Methods 0.000 claims abstract description 18
- 238000004519 manufacturing process Methods 0.000 claims description 25
- 238000012549 training Methods 0.000 claims description 17
- 230000008569 process Effects 0.000 claims description 12
- 238000007726 management method Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 8
- 238000010219 correlation analysis Methods 0.000 claims description 4
- 230000035515 penetration Effects 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000010206 sensitivity analysis Methods 0.000 claims description 4
- 238000009313 farming Methods 0.000 claims description 3
- 238000007405 data analysis Methods 0.000 abstract description 21
- 238000005516 engineering process Methods 0.000 abstract description 10
- 230000008901 benefit Effects 0.000 abstract description 6
- 230000003993 interaction Effects 0.000 abstract description 5
- 238000003860 storage Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 6
- 201000010099 disease Diseases 0.000 description 5
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000006651 lactation Effects 0.000 description 4
- 241000282887 Suidae Species 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000009395 breeding Methods 0.000 description 3
- 230000001488 breeding effect Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000000354 decomposition reaction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000012173 estrus Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000035935 pregnancy Effects 0.000 description 2
- 230000002035 prolonged effect Effects 0.000 description 2
- 238000003307 slaughter Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000012351 Integrated analysis Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 244000144972 livestock Species 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000007306 turnover Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Landscapes
- Machine Translation (AREA)
Abstract
The invention discloses a method, a device and equipment for analyzing farm data indexes based on a large language model, which comprises the following steps: receiving user input data comprising login information and query content, and determining a current farm according to the login information; analyzing the query content by combining the prompt engineering of the pre-trained large language model with the current farm to obtain a data index to be analyzed; processing the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results; and integrating and analyzing the data return results through the large language model, generating an analysis reply result, and outputting the analysis reply result to a user. By utilizing the advantages of a large language model, a complex data analysis technology is converted into a simple and easily understood operation mode, and by natural language interaction, non-professional data analysis personnel can easily get on hand, so that quick analysis and application of data are realized.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method, a device and equipment for analyzing farm data indexes based on a large language model.
Background
With the rapid development of information technology, the front-end technologies such as internet technology, big data technology, artificial intelligence and the like are widely applied to various industries. In particular in the field of data analysis, these techniques provide an effective means for data acquisition, storage, analysis and visualization. However, these techniques and products often come with high technological thresholds that make it difficult for non-professional data analysts to use them conveniently and efficiently.
In the field of farms, the collection and analysis of data have important significance in the aspects of improving production efficiency, optimizing feeding management, reducing disease incidence and the like. However, most farms still rely on traditional data analysis methods, which tend to be inefficient and difficult to address large-scale, complex data processing requirements. In addition, farm workers often lack deep knowledge of data analysis techniques, resulting in inefficient data utilization and difficulty in fully exploiting the role of data in farm management. Although there are already some data analysis products on the market with friendly interaction patterns, they tend to focus on versatility and it is difficult to fully meet specific requirements in the field of farms. Especially in combination of business knowledge in the vertical industry and various data analysis technologies, the existing products cannot achieve the balance, so that the farms still face higher technical thresholds when the data are utilized to promote business improvement or innovation.
The development of Large Language Models (LLM) provides new opportunities for reducing data usage costs and innovative data consumption approaches. However, in the field of farms, how to combine large language models with farm data analysis to achieve civilian data analysis remains a significant challenge. At present, a complex data index analysis technology is not packaged into simple and easy-to-use analysis capability, and a practical, feasible and product service close to individual requirements is provided for non-professional data analysts through natural language interaction. Therefore, how to package a complex data index analysis technology into a simple and easy-to-use analysis capability, and provide practical, feasible and product services close to individual requirements for non-professional data analysts through natural language interaction has become a problem to be solved.
Disclosure of Invention
In view of the above, the invention aims to provide a large language model-based data index analysis method, device and equipment for farms, which aim to solve the problems of how to realize convenient and efficient data index analysis service according to the demands of the farms.
In order to achieve the above object, the present invention provides a method for analyzing farm data indexes based on a large language model, the method comprising:
Receiving user input data comprising login information and query content, and determining a current farm according to the login information;
analyzing the query content by combining the prompt engineering of the pre-trained large language model with the current farm to obtain a data index to be analyzed;
Processing the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results;
And integrating and analyzing the data return results through the large language model, generating an analysis reply result, and outputting the analysis reply result to a user.
Preferably, the analyzing the query content by combining the prompt engineering of the large language model completed through pre-training with the current farm to obtain the data index to be analyzed includes:
Analyzing the query content by combining the prompt engineering of the large language model with the current farm to obtain the data index to be analyzed, wherein the data index comprises first data and second data; the first data includes farming data for the current farm, and the second data includes proper nouns of industry.
Preferably, the API interface includes an API query interface and an API analysis interface; the processing of the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results comprises the following steps:
inquiring the first data by calling the corresponding API inquiry interface to obtain an inquiry result;
Analyzing the first data and the query result by calling the corresponding API analysis interface to obtain an analysis result;
searching the second data through the preset knowledge base to obtain a search result;
And returning the query result, the analysis result and the search result as the data.
Preferably, the types of the API analysis interface include a contrast analysis API, a correlation analysis API, a prediction analysis API, a attribution analysis API, a penetration analysis API, a sensitivity analysis API, and a recommendation analysis API.
Preferably, the retrieving the second data through the preset knowledge base to obtain a retrieval result includes:
vector retrieval is carried out on proper nouns or industry terms corresponding to the second data in the preset knowledge base, and the retrieved knowledge concepts are used as the retrieval results.
Preferably, the training process of the large language model includes:
The method comprises the steps of obtaining data indexes and preprocessing the data indexes to serve as a training data set, wherein the data indexes comprise production data indexes, cost data indexes, sales data indexes, purchasing data indexes, financial data indexes and management data indexes;
And inputting the training data set into a preset large language model for pre-training to obtain the trained large language model.
Preferably, the outputting the analysis reply result to the user includes:
and outputting the analysis reply result to a user for display in a mode of analysis explanation, data source, data chart and/or reference link.
In order to achieve the above object, the present invention further provides a plant data index analysis device based on a large language model, the device comprising:
The receiving unit is used for receiving user input data comprising login information and query content and determining a current farm according to the login information;
the analysis unit is used for analyzing the query content by combining the prompt engineering of the pre-trained large language model with the current farm to obtain a data index to be analyzed;
The processing unit is used for processing the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results;
And the generating unit is used for integrating and analyzing the data return results through the large language model, generating an analysis reply result and outputting the analysis reply result to a user.
In order to achieve the above object, the present invention also proposes a large language model based farm data index analysis device, comprising a processor, a memory and a computer program stored in the memory, the computer program being executed by the processor to implement the steps of a large language model based farm data index analysis method as described in the above embodiments.
To achieve the above object, the present invention also proposes a computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of a method for analyzing farm data indicators based on a large language model as described in the above embodiments.
The beneficial effects are that:
According to the scheme, the query content of the user is analyzed based on the large language model, different API interfaces and the preset knowledge base are called to process different analyzed task branches, and the obtained multiple returned results are integrated and analyzed to generate a final analysis reply result, so that the technical threshold can be reduced, and the civilian data analysis can be realized. By utilizing the advantages of a large language model, a complex data analysis technology is converted into a simple and easily understood operation mode, and by natural language interaction, non-professional data analysis personnel can easily get on hand, so that quick analysis and application of data are realized.
According to the scheme, through combining the specific requirements of the field of farms, accurate collection, efficient processing and visual display of data can be realized, and the workers of the farms can quickly acquire key information and make accurate decisions, so that the management efficiency of the farms is improved, and unnecessary losses are reduced; by fully utilizing the collected data, the growth condition, disease occurrence rule and the like of livestock are analyzed, so that the feeding management scheme is optimized, and the method is beneficial to reducing the disease occurrence rate, improving the feed utilization rate and increasing the breeding benefit.
The technical innovation in the field of farms is facilitated, the development of industries is led, and with the deep application of big data and artificial intelligence technology, the data analysis capability of the farms is further improved, and powerful support is provided for sustainable development of the industries.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for analyzing farm data indexes based on a large language model according to an embodiment of the present invention.
Fig. 2 is an overall flow chart according to an embodiment of the invention.
Fig. 3 is a schematic structural diagram of a farm data index analysis device based on a large language model according to an embodiment of the present invention.
The realization of the object, the functional characteristics and the advantages of the invention will be further described with reference to the accompanying drawings in connection with the embodiments.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, based on the embodiments of the invention, which are apparent to those of ordinary skill in the art without inventive faculty, are intended to be within the scope of the invention. Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, based on the embodiments of the invention, which are apparent to those of ordinary skill in the art without inventive faculty, are intended to be within the scope of the invention.
In the description of the present invention, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature.
The following describes the invention in detail with reference to examples.
Referring to fig. 1, a flow chart of a method for analyzing farm data indexes based on a large language model according to an embodiment of the invention is shown.
In this embodiment, the method includes:
S11, receiving user input data comprising login information and query content, and determining the current farm according to the login information.
And S12, analyzing the query content by combining the prompt engineering of the large language model which is finished through pre-training with the current farm to obtain the data index to be analyzed.
Further, in step S12, the analyzing the query content by the prompt engineering of the large language model through pre-training in combination with the current farm to obtain a data index to be analyzed includes:
Analyzing the query content by combining the prompt engineering of the large language model with the current farm to obtain the data index to be analyzed, wherein the data index comprises first data and second data; the first data includes farming data for the current farm, and the second data includes proper nouns of industry.
Further, parsing the query content through the prompt engineering of the large language model in combination with the current farm includes splitting the query content into a plurality of tasks in a thinking chain (CoT, chain of Thoughts) manner of the large language model. Specifically:
1. Preset System Prompts (system prompt project set prompt word): "your role is { farm management data analyst }, you need { get specific farm name in that context }, and need { extract special keywords in a given sentence for subsequent query of the corresponding definition from the knowledge base }, while you need { analyze user intent, task of related data query and data analysis }.
2. Analyzing the input query content by using a thinking chain, and performing step-by-step reasoning according to the preset prompt word (the core of the CoT is analyzing the input content and performing step-by-step reasoning according to the system setting), including:
step0, perceiving keywords in the input query content, including English abbreviations, names of people, website names and product names appearing in sentences. The code is as follows:
{‘step’: step0,
‘input’: $query
' thoughts ' call system API to obtain the data of farm where user is located, solve the problem of user '
'Output' [ 'pig farm name is ×' ],
}
Step1, performing knowledge base search to obtain a structured result.
{‘step’: step1,
‘input’: $query
The prime abbreviation of ' thoughts ' non-manufacturing days is not common and requires further querying of the knowledge base to determine its meaning '
'Output': [ 'non-productive days are:' ],
}
Step2, executing a specific API to obtain data, and planning structured input and output by combining knowledge base interpretation.
{‘step’: step2,
‘input’: $query
The ' thoughts ' problem relates to data queries, which must be queried for data and simply analyzed '
'Output': result of analysis: good' ] is achieved by the fact that,
}
3. After corresponding data is obtained based on the decomposed task branches, rationality Reflection (Reflection) is carried out in combination with user intention, and specific tasks are executed after clear and reasonable planning is determined.
In this embodiment, the method is applied to a farm. The following examples are mainly illustrated by application in pig farms. Reference is made to figure 2. The method comprises the steps of receiving login information of a user entering a data analysis system and query content input by the user, decomposing the query content into a plurality of tasks which correspond to data indexes to be analyzed and comprise data query and data analysis through prompt engineering (Prompt Engineering) of a preset large language model, and generating output results according to structured input execution instructions of different task branches by calling different task branches set by a model channel.
For example, the content of the query entered by the user is "why is the number of non-production days of my pig farm so high? "; analyzing the query content through a large language model to obtain the following data:
"My pig farm" is based on the settings of the user information sub-module, and should be understood as "query pig farm name";
"days of non-production" is based on settings in the knowledge base submodule, and should be understood as "query this term";
"why it is so high" is based on the settings in the data interface sub-module, and should be understood as "query index data and interpret the cause of the high".
The analysis system comprises an input receiving module (comprising a user information sub-module, a knowledge base sub-module and a data interface sub-module), a large model planning module, a vectorization module, a data query module, an analysis task module and a large model generation module. Wherein,
The input receiving module is used for receiving natural language content input by a user through a (dialogue) interactive interface;
The large model planning module is used for decomposing input query content into input instructions, calling corresponding task branches according to the acquired user login information, model setting conditions and the like, and acquiring preset data analysis Prompt (Prompt), preset model channel setting (Pipeline) and preset task scheduling execution Prompt (Tips); a preset data analysis prompt is used for restricting the style and the sample of the large model output reply, for example, "as a data analysis expert, always using markdown to display the final analysis result", etc.; the preset model channel setting is used for selecting different large models or large model versions and generating different replies; for example, GPT-3.5-turbo or GPT-4, or GLM-2, or GLM-4V, or Qwen-turbo equi-sized models may be selected; the preset task execution prompt words are displayed to the user in a standard foreground, and the display contents of different steps can be defined, for example, certain process display or non-display can be defined, and the display contents can be defined;
The vectorization module is used for acquiring matched knowledge base contents according to the input instruction after the large model decomposition, and matching the industry knowledge base stored in the vector database with proper nouns or industry terms in the input instruction and the like in a vector calculation mode;
The data query module is used for acquiring corresponding data query results from the input instructions after the large model decomposition, and acquiring the corresponding query results by adopting different data interfaces (APIs) through different data topics; the general query question may directly call the data query interface to obtain data and reply, for example, if the user asks: when the expression problems of 'how much is currently …', 'how much is the present month', 'how much is present today', and the like are solved, the large model can call the data query module to query and reply;
the analysis task module is used for executing corresponding data analysis tasks through a plurality of data processing nodes by adopting different big data analysis models according to a preset analysis flow or an analysis flow constructed by a user;
The large model generation module is used for receiving the returned results of the vectorization module, the data query module and the analysis task module, and the incoming data of the large model planning module, and generating structured output content through a preset data interpretation prompt (Promopts), a data chart generation service (E-CHART SERVICE) and a reference link Table (REFERENCE LINK Table).
S13, processing the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results.
Further, in step S13, the API interface includes an API query interface and an API analysis interface; the processing of the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results comprises the following steps:
S13-1, inquiring the first data by calling the corresponding API inquiry interface to obtain an inquiry result;
S13-2, analyzing the first data and the query result by calling the corresponding API analysis interface to obtain an analysis result;
s13-3, searching the second data through the preset knowledge base to obtain a search result;
S13-4, taking the query result, the analysis result and the search result as the data return result.
The types of the API analysis interfaces comprise a contrast analysis API, a correlation analysis API, a prediction analysis API, an attribution analysis API, a penetration analysis API, a sensitivity analysis API and a recommendation analysis API.
Further, in step S13-3, the retrieving, through the preset knowledge base, the second data to obtain a retrieval result includes:
vector retrieval is carried out on proper nouns or industry terms corresponding to the second data in the preset knowledge base, and the retrieved knowledge concepts are used as the retrieval results.
In this embodiment, based on the above-mentioned multiple data obtained by decomposing the query content, by calling different task branches to return corresponding results, including processing the first data through an application program interface provided by a real-time digital bin, including executing a data API query interface to query a corresponding data index, obtaining a query result, and analyzing the data index by calling a corresponding API analysis interface to obtain an analysis result; and executing retrieval of the knowledge base, obtaining corresponding knowledge concepts in the knowledge base as retrieval results by calculating vector relations between related texts in the preset knowledge base and second data comprising proper nouns or industry terms through matching, and further sending the returned results to the large language model. The manner of calling the API interface comprises the steps of requesting an address: an API request address defined in the form of a URL; the request mode is as follows: the POST mode and the GET mode are included, and corresponding data results can be obtained; request parameters: including parameter name (English), data type (String, integer, long, flow, double, boolean and Date format), enumerated value (enumerate variable parameter) and default value; output parameters: i.e., the query results, include parameter names (English), data types (String, integer, long, flow, double, boolean and Date formats, etc.), and presentation names (text).
Based on the above example, the user information query interface is invoked to query the current pig farm (i.e., pig farm name) based on the login information, and replace "my pig farm" with this pig farm name in reply; searching (RETRIEVAL) through a knowledge base, inquiring the noun meaning represented by the non-production days, and subsequently giving explanation in a reply; inquiring the specific numerical value of the index of the non-production days by calling an API (application program interface) inquiry interface, and presenting the specific numerical value in a reply; and analyzing the reason of the increase of the index of the non-production days by calling an API analysis interface, and presenting the result through a chart, characters and analysis results.
The large language model performs selection of different API analysis interfaces in a fuzzy matching mode to obtain a return result closest to the problem of the user, wherein the types of the API analysis interfaces comprise:
Comparative analysis API: comparing differences between the two or more data to identify changes and trends;
correlation analysis API: exploring what association exists between variables in the data, such as Pearson correlation and the like;
Predictive analysis API: using statistical or machine learning algorithms to predict future events or trends;
Attribution analysis API: a method for determining a cause of a data change to identify factors affecting the data change and to optimize a business process;
Penetration analysis API: deep explore the association relationship between the data to identify patterns and trends hidden in the data;
sensitivity analysis API: the method is used for evaluating the influence of variables on the result so as to optimize the business strategy and decision;
Recommendation analysis API: personalized suggestions are generated based on the user behavior data to enhance the user experience and increase sales.
Continuing with the above example, the user query includes: when the expression problems such as why a certain index is raised/lowered, what causes a certain index to be raised or lowered, what causes a certain index to be lowered, and the like are solved, the pre-trained large language model calls the attribution analysis API to obtain the analysis result.
S14, integrating and analyzing the data return results through the large language model, generating analysis reply results, and outputting the analysis reply results to a user.
Further, the training process of the large language model comprises the following steps:
The method comprises the steps of obtaining data indexes and preprocessing the data indexes to serve as a training data set, wherein the data indexes comprise production data indexes, cost data indexes, sales data indexes, purchasing data indexes, financial data indexes and management data indexes;
And inputting the training data set into a preset large language model for pre-training to obtain the trained large language model.
Further, the outputting the analysis reply result to the user includes:
and outputting the analysis reply result to a user for display in a mode of analysis explanation, data source, data chart and/or reference link.
In this embodiment, the obtained data index is preprocessed and then input into a selected large language model (such as GPT-3.5-turbo or GPT-4, or GLM-2, or GLM-4V, or Qwen-turbo, etc.), and the large language model is pre-trained; the preprocessing process comprises data cleaning, data conversion (including data compression, discretization, normalization, standardization and the like), data integration, data auditing and the like, and aims to improve the accuracy, applicability, timeliness and consistency of data. The data index includes:
production data index: corresponding index data for measuring production conditions, such as factory indexes of daily throughput, month production number and the like; taking pig farm production as an example, the index can be the number of pigsty in the pigsty, the annual litter size (PSY) of the head average pig, and the like;
cost data index: corresponding index data for measuring the cost spent in the production and management process, such as monthly labor cost, monthly hydropower cost, etc.;
sales data index: index data for measuring sales conditions of enterprises, such as sales numbers of sows, sales amounts of fattening pigs and the like;
purchasing data indexes: index data for measuring the purchasing condition of enterprises, such as feed purchasing quantity, suckling pig purchasing quantity, medicine purchasing amount and the like;
Financial data index: index data for measuring financial conditions of enterprises, such as liability, fund turnover, profit margin, etc.;
and (3) operating data indexes: index data for measuring business operations, such as market share, number of users, sales income, etc.
The system can also provide personalized advice and training content for users to display the content by means of understanding data results, analyzing result interpretation, generating data charts according to data sources, giving reference links and the like through the natural language generation capability of the large language model. Continuing with the example above, obtaining includes:
Search results: the term "non-productive days, i.e., NPD index, wherein english is Non Production Days, means the number of days that the sow does not perform productive activities in the pig farm, including estrus, gestation, lactation, etc. "; query results: "the non-production days of Xinnong pig farm is 24 days"; analysis results: the main reason why the non-production days of the Xinnong pig farm are higher than the industry average 22.5 days is that the period of lactation is long (average 13.5 days) through attribution analysis of a trend, so that the breeding cycle is prolonged, and the marketing delay is caused. Through the recent operation records of pig farms, disease occurrence conditions and feeding conditions in pig farms also affect the growth rate of pigs, resulting in increased time to slaughter. "by integrating the above returned results, generating final reply content:
"based on pig networking data, the non-production days (NPD) of the Xinnong pig farm was 24 days. The non-production days refer to the days of the sow not carrying out production activities in the pig farm, including estrus, gestation, lactation and the like. The production efficiency and economic benefit of the pig farm can be influenced by the index.
The non-production days of the Xinnong pig farm are higher than the industry average 22.5 days, and the main reason is that the period of lactation is long (average 13.5 days) through attribution analysis of a period of trend, so that the breeding period is prolonged, and the marketing delay is caused. Through the recent operation records of pig farms, disease occurrence conditions and feeding conditions in pig farms also affect the growth rate of pigs, resulting in increased time to slaughter.
The above data shows that the Xinnong pig farm performs well on the NPD index, but there is some room for improvement, which means that the non-production days are longer than the industry leading enterprises. If you want, I can help you analyze against the national average index in order to better assess your pig farm production.
Referring to fig. 3, a schematic structural diagram of a farm data index analysis device based on a large language model according to an embodiment of the present invention is shown.
In this embodiment, the apparatus 20 includes:
a receiving unit 21 for receiving user input data including login information and query content, and determining a current farm based on the login information;
The parsing unit 22 is configured to parse the query content by combining the prompt engineering of the pre-trained large language model with the current farm, so as to obtain a data index to be analyzed;
The processing unit 23 is configured to process the data index to be analyzed by calling an API interface and/or a preset knowledge base, so as to obtain a plurality of data return results;
and the generating unit 24 is used for carrying out integrated analysis on a plurality of data return results through the large language model, generating analysis reply results and outputting the analysis reply results to a user.
The respective unit modules of the apparatus 20 may perform the corresponding steps in the above method embodiments, so that the detailed description of the respective unit modules is omitted herein.
The embodiment of the invention also provides a large language model-based farm data index analysis device, which comprises the large language model-based farm data index analysis device, wherein the large language model-based farm data index analysis device can adopt the structure of the embodiment of fig. 3, correspondingly, the technical scheme of the embodiment of the method shown in fig. 1 can be executed, the implementation principle and the technical effect are similar, and details can be found in related records in the embodiment, so that the details are not repeated.
The apparatus comprises: a device with a photographing function such as a mobile phone, a digital camera or a tablet computer, or a device with an image processing function, or a device with an image display function. The device may include a memory, a processor, an input unit, a display unit, a power source, and the like.
The memory may be used to store software programs and modules, and the processor executes the software programs and modules stored in the memory to perform various functional applications and data processing. The memory 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 (e.g., an image playing function, etc.) required for at least one function, and the like; the storage data area may store data created according to the use of the device, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory may also include a memory controller to provide access to the memory by the processor and the input unit.
The input unit may be used to receive input digital or character or image information and to generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control. Specifically, the input unit of the present embodiment may include a touch-sensitive surface (e.g., a touch display screen) and other input devices in addition to the camera.
The display unit may be used to display information entered by a user or provided to a user as well as various graphical user interfaces of the device, which may be composed of graphics, text, icons, video and any combination thereof. The display unit may include a display panel, and optionally, the display panel may be configured in the form of an LCD (Liquid CRYSTAL DISPLAY), an OLED (Organic Light-Emitting Diode), or the like. Further, the touch-sensitive surface may overlay the display panel, and upon detection of a touch operation thereon or thereabout, the touch-sensitive surface is communicated to the processor to determine the type of touch event, and the processor then provides a corresponding visual output on the display panel based on the type of touch event.
The embodiment of the present invention also provides a computer readable storage medium, which may be a computer readable storage medium contained in the memory in the above embodiment; or may be a computer-readable storage medium, alone, that is not assembled into a device. The computer readable storage medium has stored therein at least one instruction that is loaded and executed by a processor to implement the large language model based farm data index analysis method shown in fig. 1. The computer readable storage medium may be a read-only memory, a magnetic disk or optical disk, etc.
It should be noted that, in the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described as different from other embodiments, and identical and similar parts between the embodiments are all enough to be referred to each other. For the device embodiments, the apparatus embodiments and the storage medium embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference is made to the description of the method embodiments for relevant points.
Also, herein, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
While the foregoing description illustrates and describes the preferred embodiments of the present invention, it is to be understood that the invention is not limited to the forms disclosed herein, but is not to be construed as limited to other embodiments, but is capable of use in various other combinations, modifications and environments and is capable of changes or modifications within the scope of the inventive concept, either as described above or as a matter of skill or knowledge in the relevant art. And that modifications and variations which do not depart from the spirit and scope of the invention are intended to be within the scope of the appended claims.
Claims (10)
1. A method for analyzing farm data indexes based on a large language model, the method comprising:
Receiving user input data comprising login information and query content, and determining a current farm according to the login information;
analyzing the query content by combining the prompt engineering of the pre-trained large language model with the current farm to obtain a data index to be analyzed;
Processing the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results;
And integrating and analyzing the data return results through the large language model, generating an analysis reply result, and outputting the analysis reply result to a user.
2. The method for analyzing the data index of the farm based on the large language model according to claim 1, wherein the analyzing the query content by combining the prompt engineering of the large language model completed through the pre-training with the current farm to obtain the data index to be analyzed comprises the following steps:
Analyzing the query content by combining the prompt engineering of the large language model with the current farm to obtain the data index to be analyzed, wherein the data index comprises first data and second data; the first data includes farming data for the current farm, and the second data includes proper nouns of industry.
3. The method for analyzing farm data indexes based on the large language model according to claim 2, wherein the API interface comprises an API query interface and an API analysis interface; the processing of the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results comprises the following steps:
inquiring the first data by calling the corresponding API inquiry interface to obtain an inquiry result;
Analyzing the first data and the query result by calling the corresponding API analysis interface to obtain an analysis result;
searching the second data through the preset knowledge base to obtain a search result;
And returning the query result, the analysis result and the search result as the data.
4. A method of analyzing farm data metrics based on a large language model according to claim 3, characterized in that the types of API analysis interfaces include a contrast analysis API, a correlation analysis API, a predictive analysis API, an attribution analysis API, a penetration analysis API, a sensitivity analysis API, and a recommendation analysis API.
5. A method for analyzing a data index of a farm based on a large language model according to claim 3, wherein the retrieving the second data through the preset knowledge base to obtain a retrieval result comprises:
vector retrieval is carried out on proper nouns or industry terms corresponding to the second data in the preset knowledge base, and the retrieved knowledge concepts are used as the retrieval results.
6. The method for analyzing farm data indexes based on a large language model according to claim 1, wherein the training process of the large language model comprises the following steps:
The method comprises the steps of obtaining data indexes and preprocessing the data indexes to serve as a training data set, wherein the data indexes comprise production data indexes, cost data indexes, sales data indexes, purchasing data indexes, financial data indexes and management data indexes;
And inputting the training data set into a preset large language model for pre-training to obtain the trained large language model.
7. The method for analyzing farm data indexes based on a large language model according to claim 1, wherein the step of outputting the analysis reply result to the user comprises the steps of:
and outputting the analysis reply result to a user for display in a mode of analysis explanation, data source, data chart and/or reference link.
8. A large language model-based farm data index analysis device, the device comprising:
The receiving unit is used for receiving user input data comprising login information and query content and determining a current farm according to the login information;
the analysis unit is used for analyzing the query content by combining the prompt engineering of the pre-trained large language model with the current farm to obtain a data index to be analyzed;
The processing unit is used for processing the data index to be analyzed by calling an API interface and/or a preset knowledge base to obtain a plurality of data return results;
And the generating unit is used for integrating and analyzing the data return results through the large language model, generating an analysis reply result and outputting the analysis reply result to a user.
9. A large language model based farm data index analysis device comprising a processor, a memory and a computer program stored in the memory, the computer program being executed by the processor to carry out the steps of a large language model based farm data index analysis method according to any of claims 1 to 7.
10. A computer program product comprising computer programs/instructions which when executed by a processor implement the steps of a large language model based farm data index analysis method according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410586352.3A CN118152430A (en) | 2024-05-13 | 2024-05-13 | Large language model-based farm data index analysis method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410586352.3A CN118152430A (en) | 2024-05-13 | 2024-05-13 | Large language model-based farm data index analysis method, device and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118152430A true CN118152430A (en) | 2024-06-07 |
Family
ID=91297493
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410586352.3A Pending CN118152430A (en) | 2024-05-13 | 2024-05-13 | Large language model-based farm data index analysis method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118152430A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116821103A (en) * | 2023-08-29 | 2023-09-29 | 腾讯科技（深圳）有限公司 | Data processing method, device, equipment and computer readable storage medium |
| US11861320B1 (en) * | 2023-02-27 | 2024-01-02 | Casetext, Inc. | Text reduction and analysis interface to a text generation modeling system |
| CN117493511A (en) * | 2023-11-02 | 2024-02-02 | 国农(重庆)生猪大数据产业发展有限公司 | Automatic question-answering system and method in live pig field based on large language model |
| CN117786077A (en) * | 2023-12-26 | 2024-03-29 | 新希望六和股份有限公司 | Livestock knowledge question-answering method and device based on large language model |
| CN117992594A (en) * | 2024-02-07 | 2024-05-07 | 李扬 | User accompanying method, system, equipment and storage medium based on generation type AI |
-
2024
- 2024-05-13 CN CN202410586352.3A patent/CN118152430A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11861320B1 (en) * | 2023-02-27 | 2024-01-02 | Casetext, Inc. | Text reduction and analysis interface to a text generation modeling system |
| CN116821103A (en) * | 2023-08-29 | 2023-09-29 | 腾讯科技（深圳）有限公司 | Data processing method, device, equipment and computer readable storage medium |
| CN117493511A (en) * | 2023-11-02 | 2024-02-02 | 国农(重庆)生猪大数据产业发展有限公司 | Automatic question-answering system and method in live pig field based on large language model |
| CN117786077A (en) * | 2023-12-26 | 2024-03-29 | 新希望六和股份有限公司 | Livestock knowledge question-answering method and device based on large language model |
| CN117992594A (en) * | 2024-02-07 | 2024-05-07 | 李扬 | User accompanying method, system, equipment and storage medium based on generation type AI |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110825882B (en) | Knowledge graph-based information system management method | |
| US8028003B2 (en) | System and method for presenting survey data over a network | |
| US20050120009A1 (en) | System, method and computer program application for transforming unstructured text | |
| US11687826B2 (en) | Artificial intelligence (AI) based innovation data processing system | |
| KR102565455B1 (en) | Domain-specific language interpreter and interactive visual interface for rapid screening | |
| CN102200974A (en) | Unified information retrieval intelligent agent system and method for search engine | |
| WO2023130837A1 (en) | Automatic machine learning implementation method, platform and apparatus for scientific research application | |
| Groen et al. | Is there Really a Need for Using NLP to Elicit Requirements? A Benchmarking Study to Assess Scalability of Manual Analysis. | |
| Miao et al. | A dynamic financial knowledge graph based on reinforcement learning and transfer learning | |
| Huynh et al. | Big data analytics capabilities: Patchwork or progress? A systematic review of the status quo and implications for future research | |
| Kim et al. | Semantic computing and business intelligence | |
| CN111125145A (en) | Automatic system for acquiring database information through natural language | |
| Soni et al. | A survey on automatic dashboard recommendation systems | |
| CN117149804A (en) | Data processing method, device, electronic equipment and storage medium | |
| Meth et al. | Exploring design principles of task elicitation systems for unrestricted natural language documents | |
| CN118152430A (en) | Large language model-based farm data index analysis method, device and equipment | |
| Shahzad et al. | Towards goal-driven access to process warehouse: Integrating goals with process warehouse for business process analysis | |
| Yeo et al. | Natural language interface for process mining queries in healthcare | |
| Hu et al. | Data and information quality: an information-theoretic perspective | |
| Elshan et al. | Unveiling Challenges and Opportunities in Low Code Development Platforms: A StackOverflow Analysis | |
| Guessoum et al. | Dealing with decisional natural language why-question in business intelligence | |
| EP3633578A1 (en) | Method for forecasting sales volume | |
| Dau et al. | Formal concept analysis for qualitative data analysis over triple stores | |
| Weigand et al. | How to identify your design science research artifact | |
| Zhao et al. | A Literature Review of Literature Reviews in Pattern Analysis and Machine Intelligence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |