CN102760275B - A kind of information handling system for agriculture of city type comprehensive evaluation - Google Patents

Abstract

The invention relates to an information processing system for comprehensive evaluation of urban agriculture, comprising: an input module for obtaining evaluation indicators and corresponding evaluation values; a storage module for storing prior knowledge templates, evaluation index sets, evaluation Reference threshold and rule set and historical data; evaluation processing module, based on the basic evaluation indicators obtained by the input module and the corresponding evaluation values, as well as the prior knowledge templates, evaluation index sets, evaluation reference thresholds and rule sets, and historical data stored in the storage module Perform processing to generate an evaluation result; an output module is used to output the evaluation result generated by the evaluation processing module. Compared with the prior art, the present invention automatically constructs the index system and invokes the index algorithm, thereby dynamically and multi-level information processing on the urban agriculture index, and obtains a comprehensive score.

Description

An Information Processing System for Comprehensive Evaluation of Urban Agriculture

technical field

The invention relates to an information processing system, in particular to an information processing system for comprehensive evaluation of urban agriculture.

Background technique

For a long time, urban agriculture, as a general concept, has hindered the deepening of research and the advancement of practice to a large extent. In fact, we can divide the urban agriculture that people use in different situations into three levels. The first level, that is, the macro level, is a part of the entire urban economic and social complex, which we call the urban agricultural system; the second level, that is, the regional level, is the agriculture in a certain area of the urban agricultural system. It is called urban regional agriculture; the third level, the micro level, is the agricultural management unit with independent economic interests in the urban agricultural system, which we call urban agricultural management unit. Whether it is an agricultural system, regional agriculture or agricultural management unit, the modification of "urban type" in front of it shows that it is significantly different from the corresponding "agricultural area type" agriculture. On the one hand, this difference is formed by the natural evolution of urban-rural interaction, and on the other hand, it is also formed by conscious promotion to adapt to urban expansion and meet the needs of urban development.

The evaluation index system of urban agriculture not only represents a more comprehensive and specific understanding of urban agriculture, but also serves as an action guide to promote the healthy and rapid development of urban agriculture. Evaluation index system has always been one of the key contents of urban agriculture research. Han Shiyuan designed eight indicators including per capita GDP, scientific and technological contribution rate, and tree coverage rate to describe the comprehensive development degree and overall level of urban agriculture. Huang Yinghui designed four types of indicators, including the level of agricultural security, the level of comprehensive agricultural production, the level of agricultural ecological environment and resource utilization, and the level of agricultural social services. Culture, etc. designed 21 indicators from five aspects including comprehensive production level, social service level, ecological security level, regional harmony, and development capacity building level. Bi Ran designed five categories of indicators, including ecological environment level, agricultural mechanization level, technological innovation, social service level and urban-rural harmony level. Chen Kai designed four types of indicators: agricultural input level, agricultural sustainable development, agricultural output level and rural social development level. Guan Hailing designed three types of indicators: economic development level, social development level and ecological development level. On the one hand, the existing evaluation index system fails to highlight the characteristics of urban agriculture "relying on urban resources and serving urban needs" from the perspective of its relationship with cities. It is greatly influenced by prior knowledge and is highly subjective. On the other hand, it is limited to the evaluation at the macro level, so that its evaluation index system and methods cannot dynamically realize the comprehensive evaluation of urban agriculture from different levels, so its versatility is greatly restricted.

To sum up, not only in the agricultural system, there are significant differences between urban agriculture and agricultural area agriculture, but also in the regional level and management unit level, there are also "urban" characteristics. Secondly, various evaluation methods are relatively subjective in the selection of indicators and calculation methods. Therefore, there is a need for a general method, system, and device to intelligently select indicators, construct systems, and integrate algorithms during the evaluation process, so as to realize comprehensive evaluation of urban agriculture in a dynamic and multi-level manner.

Contents of the invention

The purpose of the present invention is to provide an information processing system for the comprehensive evaluation of urban agriculture in order to overcome the above-mentioned defects in the prior art. Information processing is carried out on the indicators of urban agriculture to obtain a comprehensive score.

The purpose of the present invention can be achieved through the following technical solutions:

An information processing system for comprehensive evaluation of urban agriculture, including:

The input module is used to obtain the evaluation index and the corresponding evaluation value;

The storage module is used to store prior knowledge templates, evaluation index sets, evaluation reference thresholds, rule sets, and historical data;

The evaluation processing module processes the basic evaluation indicators obtained by the input module and the corresponding evaluation values, as well as the prior knowledge templates, evaluation index sets, evaluation reference thresholds, rule sets, and historical data stored in the storage module, and generates evaluation results;

The output module is used to output the evaluation result generated by the evaluation processing module.

The prior knowledge template, evaluation index set, evaluation reference threshold and rule set and historical data in the storage module are stored in the relational database as formatted data conforming to the ER model.

The evaluation processing module includes:

The index preprocessing unit is used to format the input evaluation index;

An indicator intelligent recommendation unit is used to optimize the formatted evaluation indicators;

The comprehensive evaluation calculation unit generates evaluation results based on different algorithms according to the optimized evaluation index, evaluation reference threshold and rule set;

The evaluation result fusion unit is configured to fuse evaluation results based on different algorithms to generate a final evaluation result.

The index preprocessing unit includes a dimensionless processing subunit and a scale setting subunit, which are respectively used to perform dimensionless processing and scale setting processing on the evaluation index.

The indicator intelligent recommendation unit includes:

The data cleaning sub-unit is used to standardize the input data and store it in the storage module;

The mode construction subunit calls the evaluation reference rules from the storage module, and constructs the evaluation mode according to the evaluation reference rules;

The pattern recognition subunit adopts the method of combined weighted scoring to supplement and optimize the evaluation mode constructed by the pattern construction subunit;

The indicator recommendation subunit is used for intelligent recommendation of indicators and optimization of the indicator system.

The output module includes a quantitative output unit, a graphic output unit and a predefined report format output unit.

Compared with the prior art, the present invention has the following advantages:

1) The present invention can recommend similar templates and provide guides to complete the construction of the evaluation index system based on the basic evaluation goals input by the user and prior knowledge.

2) The present invention can automatically complete the data preprocessing operations of dimensionless processing of user evaluation indicators and scale setting.

3) The present invention can intelligentize the evaluation index based on the combined weighted scoring method, complement and optimize the user's evaluation index system.

4) The present invention can provide a variety of algorithm packages for calling during the comprehensive evaluation process, and retain the interface of the extended algorithm package.

5) The present invention can dynamically fuse the comprehensive evaluation results calculated by different algorithm packages.

6) The present invention can output and visualize the evaluation results in various forms.

Description of drawings

Fig. 1 is an overall structural block diagram of the present invention;

Fig. 2 is the information processing process figure of the present invention;

Fig. 3 is an internal structural diagram of the input module;

Fig. 4 is an internal structure diagram of the indicator preprocessing unit;

Fig. 5 is an internal structure diagram of an index intelligent recommendation unit;

Fig. 6 is a schematic structural diagram of an evaluation result fusion unit;

Fig. 7 is a schematic diagram of evaluation target setting;

Figure 8 is a schematic diagram of the definition method of the custom indicator input unit;

Fig. 9 is a pattern construction flow chart of the pattern construction subunit.

Detailed ways

The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

Example

As shown in FIG. 1 , an information processing system for comprehensive evaluation of urban agriculture includes an input module 101 , a storage module 102 , an evaluation processing module 103 and an output module 104 .

Wherein, the input module 101 is used to obtain evaluation indexes and corresponding evaluation values, including definitions and values of various evaluation indexes for urban agriculture.

The storage module 102 stores structured data such as prior knowledge templates, candidate evaluation index sets, evaluation reference thresholds, and rule sets, and saves historical information and intermediate information of urban agriculture evaluation calculations during system operation. In this embodiment, the historical information of urban agriculture evaluation indicators stored in the storage module 102 is obtained on the basis of historical accumulation, and serves as the data basis for template generation, index recommendation and algorithm processing.

The evaluation processing module 103 performs index preprocessing (including dimensionless processing and scale setting) according to the evaluation index input by the input module 101, and performs evaluation index evaluation based on the user's evaluation target and the historical evaluation records stored in the storage unit 102. Intelligent recommendation processing, and algorithm calculation according to the above-mentioned two-step merged and formatted index set after cleaning, the evaluation processing module 103 presets a variety of algorithms to choose from, and finally outputs the calculation results to the output module 104 after fusion processing , which mainly includes several units as shown in Figure 2, which are:

1. The indicator preprocessing unit 202 for formatting the input evaluation indicators, including two main operations: dimensionless processing of indicators and scale setting. Dimensionless processing refers to converting the original data of all indicators into Unified scalar data without units. Scale setting means that after setting different spans of different index intervals and assigning the same score span, the average scoring method is used in each small index interval to make the scale setting more accurate. science.

2. The indicator intelligent recommendation unit 203 for optimizing the formatted evaluation indicators, supplements and optimizes the preprocessed evaluation indicator system based on the combined weighted scoring method, making the evaluation results more objective and comprehensive . At the same time, it solves the sparsity problem, cold start problem and scalability problem existing in the traditional recommendation method.

3. A comprehensive evaluation calculation unit 204 that generates evaluation results based on different algorithms according to the optimized evaluation index, evaluation reference threshold and rule set.

4. An evaluation result fusion unit 205 for fusing evaluation results based on different algorithms, as shown in FIG. 6 , to generate a final evaluation result.

The output module 104 includes a quantitative output unit, a graphical output unit, and a predefined report format output unit, and outputs the final evaluation result to the user according to the calculation result of the processing unit 103. The evaluation result can be quantified, graphically, and evaluated. output in various formats.

The specific structure of the input module 101 is shown in FIG. 3 , an evaluation target setting unit 301 , a custom index input unit 302 , a custom threshold input unit 303 , a priori template judgment unit 304 , a template guide unit 305 and an evaluation index construction unit 306 .

The evaluation target setting unit 301 is used to receive a set of top-level design concepts such as evaluation targets, principles, system frameworks, and thematic fields from users. The evaluation target setting in this embodiment is shown in FIG. 7 . The user-defined index input unit 302 is used to receive the user-defined index. The input of the index is carried out according to the tree structure, which is convenient for the user to maintain and maintain the internal reference of the data. The specific definition method is shown in FIG. 8 .

The custom threshold input unit 303 is used to receive the custom reference threshold input by the user, taking the use of solar energy and wind energy as an example:

Indicator: Proportion of solar and wind energy utilization (C15)

Metric Threshold: 5%

Threshold description: The European Union stipulates that the use of renewable energy is a mandatory indicator. It stipulates that the proportion of renewable energy in the UK will reach 20% in 2020, and that in Germany will be 18%. Japan is an energy-saving and efficient country. %. In Denmark in 2000, only the proportion of wind power generation has reached 10%, and the proportion of renewable energy use has far exceeded 20%. But generally speaking, the proportion of renewable energy used in most western countries is below 10%, while China's background value is relatively small, below 0.5%.

The a priori template judging unit 304 is used to analyze the input of the evaluation target setting unit 301, check whether there is a system framework with high similarity in the data center unit, and if there is such a template, call the template wizard unit 304 to assist the user to complete the evaluation index and the input of the threshold value, complete the rating index construction process, and output the result to the evaluation index construction unit 306 for subsequent evaluation processing.

FIG. 4 shows a schematic diagram of the internal structure of the indicator preprocessing unit in the evaluation processing module 103 of the present invention, including: an indicator dimensionless processing subunit 401 and a scale setting subunit 402 .

The dimensionless processing of indicators refers to converting the raw data of all indicators into unified scalar data without units. The index dimensionless processing subunit 401 in the present invention adopts the threshold (min-max) range average scoring method, namely:

When the indicator means that the bigger the better: $P_i=\frac{c_i-c_{\min }}{c_{\max }-c_{\min }}$

When the index value is smaller, the better: $P_i=\frac{c_{\max }-c_i}{c_{\max }-c_{\min }}$

Among them: P _i is the score after dimensionless, _ci is the original data of the index, c _max and c _min are the minimum value and maximum value of the index range respectively.

In fact, there are important flaws in the full-range average scoring method: for the same index span, the price and difficulty paid are generally different in different index intervals. For example: the difficulty of upgrading water quality from 5 to 4 is far less than that of upgrading from 3 to 2; the difficulty of increasing the profit rate from 5% to 10% is less than that of increasing from 20% to 25%. If the average scoring method is used, It is bound to have the defect of different difficulty and the same score, and the difficulty and importance represented by each point in the final total score are inconsistent, which loses the meaning of comprehensive evaluation. To avoid this situation, the scale setting subunit 402 of the present invention sets different scales for each index, and different scales endow different scoring ranges, and the average scoring method is used within the scales. For example: the value range of each indicator is divided into 5 levels (or other levels can be customized in the system according to user needs), and the scores are 0~20, 20~40, 40~60, 60~80, 80~100, (Considering that most indicators are continuous indicators, the above scoring intervals are also continuous). The non-range average setting scale will effectively avoid the above-mentioned problems, and bring greater degrees of freedom to the specific analysis of specific problems. For example, for a certain percentage indicator, its scale and scoring may be set as: 0-5% (0-20), 5%-8% (20-40), 8%-10% (40-60), 10% %~11% (60~80), 11%~11.5% (80~100). After setting different spans of different index intervals and assigning the same score span, the average scoring method is used in each small index interval, such as 5%~8% (20~40) in the above example, that is, in the 3% span indicators share an average score of 20 points. With respect to the average scoring method of the range in the general threshold, the method used in the present invention is called the stage average scoring method in the threshold, and the specific calculation formula is as follows:

When the index is bigger, the better: $P=a+20\×\frac{c_i-c_{d-\min }}{c_{d-\max }-c_{d-\min }}(a=\mathrm{0,20,40,60,80})$

When the index is smaller, the better: $P=a+20\×\frac{c_{d-\max }-c_i}{c_{d-\max }-c_{d-\min }}(a=\mathrm{0,20,40,60,80})$

Among them: P _i is the index value after dimensionless, _ci is the original data of the index, c _d-max and c _d-min are the minimum value and maximum value in each index interval respectively.

Table 1 shows the specific calculation results of the index thresholds and scoring ranges of the selected indicators:

Table 1

Fig. 5 shows the internal structure of an index intelligent recommendation unit according to an embodiment of the present invention, including: a data cleaning subunit 501, a pattern construction subunit 502, a pattern recognition subunit 503 and an index recommendation subunit 504,

Among them, the data cleaning subunit 501 is to extract relevant data from the user database, log database, and behavior database of the storage module, and after the data mining preprocessing technology, then store the standardized data in the database, the mode construction in the present invention Subunit 502 selects user-indicator relationship matrix and other related data combinations, user-indicator scoring matrix, and user search keywords and other evaluation reference rules to describe user preferences. The process of model construction is shown in Figure 9.

The pattern recognition subunit 503 of the present invention overcomes the following shortcomings of traditional collaborative filtering recommendations:

(1) Sparsity problem: In the early stage of system use or with the increase of resources in the data center, the indicators and user scores are asymmetrical, then the system does not have enough samples to calculate similar users, resulting in the obtained neighbor data being inconsistent. reliable.

(2) Cold start problem; users have no historical scoring records and access records; or new indicators have not been scored yet; these two situations will also lead to the situation that cannot be recommended.

(3) Scalability issues; as the number of users and index resources increase, the increase in data volume will lead to a decrease in system performance.

In order to make up for these shortcomings, the present invention adopts a method based on combined weighted scoring when implemented in the pattern recognition subunit 503, which can basically alleviate the above-mentioned shortcomings. The main function of the combined weighted score is to comprehensively process the average weighted scores of the rows and columns of the score matrix and calculate the predicted score, so that each user has a score value for each indicator, thereby alleviating the sparsity problem.

The combined weighted score is composed of two parts, the average weighted score of users and the average weighted score of indicators. The present invention makes the following improvements to the algorithm:

(1) First, combine the current user's existing score mean with the score deviation mean of other users' unrated indicators to obtain the current user's unrated indicator score estimate, that is, obtain the user's average weighted score for the unrated indicators;

(2) Again, recalculate the new rating average of the current user;

(3) Then, according to the existing score mean of the indicator, combined with the deviation mean between the existing score of the user with missing score and the new score mean of the user, the average weighted score of the missing indicator is obtained;

(4) Finally, combine (1) and (3) to obtain the final combined weighted score;

(5) Use the value obtained in (4) to fill in the missing values of the corresponding scoring indicators in the user-indicator scoring matrix.

The improved combined weighted scoring algorithm steps in the present invention are as follows:

Stepl: Estimate the rating of the current user u for its unrated indicator i:

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; +</span>\frac{{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; K</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; K</span>}}$

In the above formula, Indicates the average score (raw data) of all rating indicators of user u; K indicates the total number of users who score indicator i (raw data); r _k,i is the rating value of other user K for user u who has not rated indicator i; Indicates the average score (raw data) of user K for its rated indicators.

This formula estimates the possible average value of user u for unrated index i based on the mean value of the deviation between the ratings of other users K who have rated index i and their average rating deviation. Thus, the average weighted rating r _u of user u for unrated index i is obtained.

Step2: Calculate the average weighted score r _i of index i with missing score:

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; +</span>\frac{{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; unnew</span>}}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}$

In the above formula, Indicates the average score (raw data) of indicator i with missing scores; Indicates the average score of user u on all scoring indicators (including the updated data after step1); Q indicates the total number of indicator scores of user u (original data); r _u,q indicates the score value of user u on its rated indicator q .

This formula estimates the average weighted score of user u's unrated indicator i based on the mean value of the deviation between the score of user u's rated indicator and its updated score mean.

Step3: Calculate the combined weighted score r _u,i :

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\sqrt{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; =</span>\sqrt{<span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; K</span>}}\frac{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; K</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; Q</span>}}\frac{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; |</span>}$

r _u,i means using the combined weighted scoring method to obtain the estimated value of user u's rating for its unrated indicator i. In the above formula, the combined weighted score of the unrated item in the user-score matrix can be calculated, and the missing value of the corresponding item can be filled; thus, any user in the user-score matrix has a score for the index i, so it can be based on the user - The scoring matrix calculates the index similarity and searches for the nearest neighbor index set of the target index.

In the present embodiment, the specific calculation using this algorithm is shown in the following table, wherein Table 2 is the original data table, Table 3 is the table of r _u value obtained by the ratings of other users in Step1, and Table 4 is the table of other users' evaluation of other users in Step2. The score of the index gets the table of r _i values, and Table 5 is (d) Step3 to get the final filled data table.

Table 2

table 3

Table 4

table 5

Taking the missing data filling of U ₁ index score as an example, the specific steps are:

Step1: Calculate the user average weighted rating of I ₄

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; +</span>\frac{{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; K</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; K</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4.25</span>}<span\; class="notranslate">\; +</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2.33</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1.6</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 3.785</span>}$

Similarly, calculate the user average weighted rating of I ₅ ; and get U ₁ 's

Step2: Calculate the average weighted score of index i with score missing

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; Q</span>}}\frac{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1.5</span>}<span\; class="notranslate">\; +</span>\frac{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4.3117</span>}<span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1.4383</span>}$

Step3: Calculate the combined weighted score, and finally use this value to fill in the corresponding missing values in the user-score matrix.

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\sqrt{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; \&CenterDot;</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; =</span>\sqrt{<span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; K</span>}}\frac{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; K</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; \&CenterDot;</span><span\; class="notranslate">\; |</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; Q</span>}}\frac{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}<span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 3.785</span>}<span\; class="notranslate">\; \&times;</span>\mathrm{<span\; class="notranslate">\; 1.4383</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2.3332</span>}$

The present invention combines the TF-IDF algorithm to calculate the weight value for each characteristic index of the evaluation target, so that the evaluation target d can be expressed as an n-dimensional vector d=d{(k ₁ ,r ₁ ),(k ₂ ,r ₂ ),...(k _n ,r _n )}. With this vector, the following cosine similarity formula can be used to calculate the similarity between any two evaluation targets.

$\mathrm{<span\; class="notranslate">\; sim</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span>}{\sqrt{\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&times;</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}}$

The indicator recommendation subunit 504 performs intelligent recommendation of indicators and optimization of the indicator system according to the above calculation results according to the user's comprehensive evaluation requirements.

As shown in FIG. 2 , the evaluation result fusion unit 205 can receive the evaluation results obtained by a single or multiple algorithms of the comprehensive evaluation calculation unit 204 , and if the input is the result obtained by multiple algorithms, the evaluation result fusion unit 205 will be triggered to work. Taking 5 indicators as an example, the indicator evaluation results obtained by calling the algorithm package 1 for these 5 indicators are:

P ₁ =(0.149,0.157,0.374,0.124,0.192)

Similarly, the index evaluation results obtained through algorithm package 2 are:

P ₂ =(0.249,0.258,0.264,0.035,0.194)

If the average value method is used for fusion, the comprehensive evaluation result of the index is: P=(P ₁ +P ₂ )/2

P=(0.20,0.21,0.32,0.08,0.20)

The fusion of evaluation results in the present invention is not limited to the average value method, and other methods can be used according to specific evaluation needs.

The data fused with the evaluation results is input into the output unit 104, and finally output in various forms such as quantification, graphics, and evaluation reports to complete the systematic visual expression of the comprehensive evaluation results.

The present invention automatically constructs an index system and invokes an index algorithm, thereby dynamically and multi-level information processing on the urban agriculture index, and obtains a comprehensive score.

Claims (4)

1. for an information handling system for agriculture of city type comprehensive evaluation, it is characterized in that, comprising:

Load module, for obtaining evaluation index and corresponding figure of merit;

Memory module, for preserving priori template, evaluation index set, evaluation reference threshold value and regular collection and historical data;

Evaluate processing module, priori template, evaluation index set, evaluation reference threshold value and regular collection, historical data that the Basic Evaluation index obtained according to load module is preserved with corresponding figure of merit and memory module process, generation evaluation result;

Output module, for exporting the evaluation result evaluated processing module and generate;

Described evaluation processing module comprises:

Index pretreatment unit, for carrying out format process to the evaluation index of input;

Index intelligent recommendation unit, for being optimized the evaluation index after format process;

Comprehensive evaluation computing unit, generates the evaluation result based on algorithms of different according to evaluation index, evaluation reference threshold value and the regular collection after optimization;

Evaluation result integrated unit, for merging the evaluation result based on algorithms of different, generates final evaluation result;

Described index intelligent recommendation unit comprises:

Data cleansing subelement, preserves in memory module after the data of input are carried out standardization processing;

Mode construction subelement, calls evaluation reference rule from memory module, and builds evaluation model according to evaluation reference rule;

Pattern-recognition subelement, adopts the method for combined weighted scoring to carry out supplementing optimization to the evaluation model constructed by mode construction subelement;

Index recommends subelement, for the optimization of the intelligent recommendation and index system that carry out index.

2. a kind of information handling system for agriculture of city type comprehensive evaluation according to claim 1, it is characterized in that, the priori template in described memory module, evaluation index set, evaluation reference threshold value and regular collection and historical data are stored in relevant database with the formatted data meeting ER model.

3. a kind of information handling system for agriculture of city type comprehensive evaluation according to claim 1, it is characterized in that, described index pretreatment unit comprises nondimensionalization process subelement and scale arranges subelement, is respectively used to carry out nondimensionalization process and scale set handling to evaluation index.

4. a kind of information handling system for agriculture of city type comprehensive evaluation according to claim 1, is characterized in that, described output module comprises quantification output unit, Graphical output unit and predefine reporting format output unit.