CN107887026A - A kind of assembly type cancer intelligence Mapping System and method based on environmental hazard key element - Google Patents
A kind of assembly type cancer intelligence Mapping System and method based on environmental hazard key element Download PDFInfo
- Publication number
- CN107887026A CN107887026A CN201711052882.6A CN201711052882A CN107887026A CN 107887026 A CN107887026 A CN 107887026A CN 201711052882 A CN201711052882 A CN 201711052882A CN 107887026 A CN107887026 A CN 107887026A
- Authority
- CN
- China
- Prior art keywords
- cancer
- distribution
- monitoring data
- environmental risk
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 206010028980 Neoplasm Diseases 0.000 title claims abstract description 364
- 201000011510 cancer Diseases 0.000 title claims abstract description 355
- 230000007613 environmental effect Effects 0.000 title claims abstract description 142
- 238000013507 mapping Methods 0.000 title claims abstract description 96
- 238000000034 method Methods 0.000 title claims abstract description 73
- 238000009826 distribution Methods 0.000 claims abstract description 182
- 238000012216 screening Methods 0.000 claims abstract description 17
- 238000012544 monitoring process Methods 0.000 claims description 195
- 238000005192 partition Methods 0.000 claims description 115
- 230000034994 death Effects 0.000 claims description 51
- 231100000517 death Toxicity 0.000 claims description 51
- 238000000638 solvent extraction Methods 0.000 claims description 43
- 230000002776 aggregation Effects 0.000 claims description 41
- 238000004220 aggregation Methods 0.000 claims description 41
- 230000000875 corresponding effect Effects 0.000 claims description 21
- 230000008506 pathogenesis Effects 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 11
- 238000002790 cross-validation Methods 0.000 claims description 9
- 230000036541 health Effects 0.000 claims description 8
- 230000014509 gene expression Effects 0.000 claims description 7
- 238000010219 correlation analysis Methods 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 5
- 201000010099 disease Diseases 0.000 claims description 5
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 5
- 238000007689 inspection Methods 0.000 claims description 4
- 238000013316 zoning Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 claims description 3
- 238000012417 linear regression Methods 0.000 claims description 3
- 230000002596 correlated effect Effects 0.000 claims description 2
- 238000000059 patterning Methods 0.000 claims 6
- 108010063499 Sigma Factor Proteins 0.000 claims 1
- 238000012795 verification Methods 0.000 claims 1
- 206010058467 Lung neoplasm malignant Diseases 0.000 description 40
- 201000005202 lung cancer Diseases 0.000 description 40
- 208000020816 lung neoplasm Diseases 0.000 description 40
- 230000008569 process Effects 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000004075 alteration Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000003950 pathogenic mechanism Effects 0.000 description 2
- 230000002265 prevention Effects 0.000 description 2
- 208000000461 Esophageal Neoplasms Diseases 0.000 description 1
- 206010030155 Oesophageal carcinoma Diseases 0.000 description 1
- 208000015634 Rectal Neoplasms Diseases 0.000 description 1
- 208000005718 Stomach Neoplasms Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000035622 drinking Effects 0.000 description 1
- 201000004101 esophageal cancer Diseases 0.000 description 1
- 235000020983 fruit intake Nutrition 0.000 description 1
- 206010017758 gastric cancer Diseases 0.000 description 1
- 201000007270 liver cancer Diseases 0.000 description 1
- 208000014018 liver neoplasm Diseases 0.000 description 1
- 235000021274 meat intake Nutrition 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 206010038038 rectal cancer Diseases 0.000 description 1
- 201000001275 rectum cancer Diseases 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000000391 smoking effect Effects 0.000 description 1
- 201000011549 stomach cancer Diseases 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 235000001366 vegetable intake Nutrition 0.000 description 1
Landscapes
- Investigating Or Analysing Biological Materials (AREA)
- Medicines Containing Antibodies Or Antigens For Use As Internal Diagnostic Agents (AREA)
Abstract
The present invention relates to a kind of assembly type cancer intelligence Mapping System and method based on environmental hazard key element, Mapping System includes cartographic data and screens and build library module, region division module, cartographic model selecting module, cancer distribution drawing module.Drafting method carries out charting environment danger key element and data screening first, establishes cartographic data base；Then selected by user, to needing mapland region subregion, or direct not subregion；Then optimal cancer space mapping method is determined；After finally utilizing the optimal cancer space mapping method chosen, selected according to user, generation needs the cancer distribution map and Drawing Error distribution map of mapland.The present invention can effectively solve the unicity of traditional cancer distribution drawing pattern for being based only upon cancer survey data, inaccuracy etc., supplemented and optimized available for China's conventional cancer drawing pattern, solve the problems, such as that existing cancer distribution map lacks unified drafting standard, promote cancer drawing to develop towards the direction of standardization.
Description
Technical Field
The invention relates to an intelligent cancer mapping system and method, in particular to an intelligent modular cancer mapping system and method based on environmental risk factors.
Background
Cancer is a serious disease that seriously threatens human health and social development. According to the 2012 world cancer report, the newly added cancer cases in china account for about 20% of the world in 2012, the cancer death cases account for about 25% of the world in 2012, 306 ten thousand of the newly added cancer cases in china in 2012, and 220 ten thousand of the cancer deaths. The cancers with the highest incidence are lung cancer, gastric cancer, liver cancer, rectal cancer and esophageal cancer in turn. It is expected that by 2020, new cancer and death cases will reach 388 million and 276 million each year in china. The results of the third cause of death review sampling survey in the country in 2004-2005 showed that cancer death accounts for the second place of the total cause of death, accounting for 22.3% of all deaths in China; cancer has become the first cause of death in urban areas. China is wide in regions, and natural geographic environments and resident living habits have characteristics respectively, so that the cancer distribution conditions of various regions also have differences. The geographical distribution condition of main cancers of people in China is accurately estimated based on daily cancer monitoring data, and the comprehensive cancer prevention and treatment research and practice can be developed for government optimized prevention and treatment resources, health administration departments and medical and related institutions, so that basic reference information is provided. In the cancer mapping process, a number of challenges are inevitably encountered, such as: how to accurately find environmental risk elements affecting the spatial distribution of cancer morbidity or mortality, how to accurately establish mathematical relational expressions between the environmental risk elements and cancer morbidity or mortality monitoring data, how to efficiently and simultaneously carry out detailed mapping of cancers in a plurality of regions, and the like. In many cancer mapping methods at home and abroad, the selection of environmental risk factors and the mapping of cancers are realized by establishing a linear or nonlinear relation model or rule between the environmental risk factors and monitoring data of cancer morbidity or mortality. However, the existing cancer mapping methods have a certain application range, and the mapping precision of the method is influenced by a plurality of aspects such as the spatial distribution characteristics of the cancer, the representativeness of monitoring data, the influence mechanism of environmental risk factors and the like. Moreover, the pathogenesis of different types of cancers, environmental risk factors, risk groups and the like in different regions are very different, so that it is difficult to reflect the spatial distribution pattern and the law of the cancer morbidity or mortality in a plurality of regions by a single method. In addition, the existing cancer data produced by the prior art lack uniform drawing standards, and the randomness is large no matter the data is a projection and coordinate system, or the types of map special subjects and symbols.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the invention provides a modular cancer intelligent mapping system and method integrating functions of environmental risk element selection, partition selection and mapping model selection, which can be used for rapidly, intelligently and accurately mapping the spatial distribution map of the morbidity or mortality of cancer based on daily cancer monitoring data and by combining elements such as social economy, medical health, geographic environment and the like, and generating an estimation error distribution map, thereby effectively solving the problems of singleness, inaccuracy and the like of the traditional cancer distribution mapping mode based on the cancer monitoring data. In addition, the invention determines the structure and layer matching scheme, legend system and expression mode of the unified cancer space distribution map, supplements and optimizes the traditional cancer mapping mode in China, solves the problem that the existing cancer distribution map lacks unified mapping standard, and promotes the cancer mapping to develop towards the direction of standardization.
The technical solution of the invention is as follows: an intelligent mapping system and method for cancer based on environmental risk factors, which make full use of socioeconomic and geographic environmental risk factors related to cancer to estimate the incidence/rate or death/rate distribution of cancer.
The invention provides a modular cancer intelligent mapping system based on environmental risk factors, which comprises: the system comprises a mapping data screening and database building module, a geographical region module, a mapping model selection module and a cancer distribution mapping module. Wherein,
drawing data screening and library building module: selecting environmental risk factors related to the morbidity or mortality risk of the cancer to be mapped from social and economic factors, geographic environment factors and medical and health factors of the area to be mapped; fusing the cancer morbidity or mortality monitoring data and the selected environmental risk factor data according to the cancer types, deleting the data of noise, vacancy and logic error in the data, and establishing a cancer mapping database; finally, determining the environmental risk elements required by the final cancer mapping according to the correlation between the cancer morbidity or mortality monitoring data and the environmental risk element data in the cancer mapping database and whether the various data have the same spatial distribution pattern, namely whether the spatial consistency analysis exists;
a geographical partitioning module: for partitioning the region to be mapped based on the cancer pathogenesis and/or cancer data distribution; the method comprises the following steps of carrying out partitioning based on a cancer pathogenesis, namely partitioning a region to be mapped according to different subtype regions of environmental risk elements from environmental risk element distribution related to cancer morbidity or mortality; partitioning is carried out based on cancer data distribution, namely, according to the numerical difference of the cancer monitoring data shown between different subtype regions of the environmental risk elements which are irrelevant to the risks of cancer morbidity or mortality and have consistent spatial distribution, the environmental risk element partitioning mode corresponding to the maximum numerical difference of the cancer monitoring data between the subtype regions is selected to partition the region to be mapped; if the user selects a plurality of partition modes, the regional partition module can autonomously compare numerical differences of the cancer morbidity or mortality monitoring data in the subtype areas under different partition modes, and determine the partition mode with the largest difference as the final regional partition mode of the cancer mapping; finally, the user obtains a plurality of subtype areas in the geographical partitioning module; when the regional division module is selected to be skipped, the drawing system defaults to drawing the region to be drawn under the condition of no division;
a charting model selection module: the mapping module is used for selecting different cancer mapping models for each subtype area in a regionalized mode or the whole area needing mapping under the condition of no partition, and recommending the cancer default optimal mapping models of each subtype area or the whole area needing mapping in the regionalized mode to a user by integrating the spatial distribution characteristics of various cancer monitoring data and the influence of environmental risk factors determined in the mapping data screening and library building module on the cancer morbidity or mortality risk; the model library of the drawing model selection module comprises a plurality of drawing models suitable for different spatial distribution characteristics and environmental risk factor influence forms;
a cancer distribution mapping module: after the charting model selection module selects a proper cancer charting model, the cancer distribution charting module generates a distribution map of the incidence/rate or death/rate of the cancer and a charting error distribution map of the area to be charted according to the selection of a user.
Correspondingly, the invention also provides a mapping method of the modular cancer intelligent mapping system based on the environmental risk factors, which comprises the following steps:
screening and establishing a mapping data library, namely selecting environmental risk factors possibly related to the morbidity or mortality risk of the cancer to be mapped from social and economic factors, geographic environment factors and medical and health factors of a region to be mapped; fusing and denoising the cancer morbidity or mortality monitoring data and the selected environmental risk factor data, thereby establishing a cancer mapping database; finally, determining environmental risk factors required by cancer mapping according to the correlation between the cancer morbidity or mortality monitoring data and the environmental risk factor data in the cancer mapping database and whether consistent spatial distribution pattern analysis exists in various data;
a regional division step, namely dividing a region to be mapped based on a cancer pathogenesis and/or cancer data distribution; the method comprises the following steps of carrying out partitioning based on a cancer pathogenesis mechanism, namely partitioning a region to be mapped according to different subtype regions of environmental risk elements from environmental risk element distribution related to cancer morbidity or mortality; partitioning is carried out based on cancer data distribution, namely, according to numerical differences expressed by cancer monitoring data among different subtype zones of the environmental risk elements which are irrelevant to the risks of cancer onset or death but have consistent spatial distribution, the environmental risk element partitioning mode corresponding to the maximum difference is selected to partition the area to be mapped; when a plurality of partition modes are selected, the numerical difference of the cancer morbidity or mortality monitoring data in different partition modes among all subtype areas is automatically compared, and the partition mode with the largest difference is determined as the final regional partition mode of the cancer mapping;
the method comprises the following steps of (1) selecting a charting model, namely integrating the spatial distribution characteristics of various cancer monitoring data and the influence of environmental risk factors determined in a charting data screening and database building module on the cancer morbidity or mortality risk, and selecting a default optimal cancer charting model for each subtype area under a regional partition mode or the whole area needing charting under the condition of no partition based on a model database comprising a large number of charting models suitable for different spatial distribution characteristics and environmental risk factor influence forms;
and a step of mapping the cancer distribution, namely selecting a proper cancer mapping model, and generating a distribution map of the cancer morbidity or mortality of a region to be mapped and a mapping error distribution map according to the selection of a user.
The mapping method of the modular cancer intelligent mapping system based on the environmental risk factors further comprises the following preferential technical characteristics:
the determination step of the sub-type area optimal charting model comprises the following steps: after the subareas are determined through the geographical division step, the charting model selection step evaluates the influence of the spatial distribution characteristics of the cancer monitoring data and the environmental risk factors on the cancer morbidity or mortality risk in each subtype area; and then, a drawing model selection step automatically selects different drawing models suitable for the space distribution characteristics and the environment risk factor influence form from a model library, respectively draws, performs precision evaluation through cross validation, and finally uses a method with the highest precision as the optimal drawing model in the subtype area.
Wherein the step of generating the cancer profile comprises: generating the distribution data of the number of diseases/rate or the number of deaths/rate of the cancers to be mapped of different crowds such as city/county, male/female or crowds of different age groups in the area to be mapped, or generating the distribution data of the number of diseases/rate or the number of deaths/rate of the cancers of the whole crowd in the area to be mapped according to the crowd composition ratio by using the cancer distribution data results of different crowds, and simultaneously generating mapping error data by using a cross validation method in the cancer distribution mapping step; determining the matching scheme, legend and expression mode of the map layer of the cancer distribution map, representing administrative units or regional boundary ground object elements by certain different dotted line symbols or characters, representing the cancer morbidity/mortality, the cancer morbidity of relative errors and the grade of the mortality by adopting chromatic aberration, drawing the cancer morbidity/mortality or mortality/distribution map and the drawing error distribution map, and finally printing the map after the quality inspection is finished.
The steps for carrying out the partition based on the cancer pathogenesis are as follows:
determining environmental risk factors related to the risk of cancer onset or death through correlation analysis;
partitioning each screened environmental risk element according to the principle that the difference of the numerical values of the cancer monitoring data in the same subtype area is minimum and the difference between different subtype areas is maximum; firstly, setting the number of the expected subareas of the environmental risk elements; then randomly dividing the corresponding number of dangerous element value intervals as initial partitions, and calculating the value difference q between the cancer monitoring data in subtype areas corresponding to all the value intervals under the initial partitions;
wherein N in the formula (1) represents the number of all cancer monitoring points in the whole region to be mapped, N h The number of cancer monitoring points in a subtype region h (h =1, \8230;, L) of an environmental risk element, L being the number of divisions of the environmental risk element, σ 2 Is the wholeThe magnitude of the numerical difference, σ, between all cancer monitoring data for the region to be mapped h 2 Is the magnitude of the numerical difference between cancer monitoring data within subtype h region, q ∈ [0,1]The larger q is, the larger the numerical difference between the cancer monitoring data in different subtype areas is, and the better the zoning effect is; then, random numbers are added on the basis of the boundary line value of each numerical value interval of the initial partition respectively to carry out partition adjustment, and the change of the q value is calculated; if the q value under the adjusted partition is larger than the q value of the initial partition, taking the adjusted partition as the current optimal partition, and then respectively adding random numbers on the basis of the boundary line value of each numerical value interval of the current optimal partition to perform partition adjustment again; and the classification adjustment is circulated for multiple times until the adjustment times reach the set iteration times. Finally, the optimal partition with the maximum q value is used as a partition result of the element;
and (3) comparing the numerical difference values q among the cancer monitoring data in all the selected type subareas of the environmental risk elements, and taking a certain environmental risk element subarea corresponding to the largest q value as a final subarea mode based on the cancer pathogenesis.
Wherein the step of partitioning based on the cancer data distribution is as follows:
selecting environmental risk elements which are irrelevant to cancer morbidity or mortality and have relatively consistent spatial distribution, and performing spatial superposition with cancer monitoring data one by one, thereby realizing spatial partitioning of the cancer monitoring data based on each selected environmental risk element;
and (2) comparing the numerical difference values q between the cancer monitoring data in all the subtype areas of the selected environmental risk elements, and taking a certain environmental risk element partition corresponding to the maximum q value as a final partition mode based on the cancer data distribution.
The step of selecting the drawing model further comprises the following steps:
the method comprises the following steps of (1) directly evaluating the influence of the spatial distribution characteristics of cancer monitoring data and environmental risk factors on the cancer morbidity or mortality risk in the whole region to be mapped if a user selects to skip the regional division; otherwise, according to the partitioning result in the regional partitioning step, respectively evaluating the influence of the spatial distribution characteristics of the cancer monitoring data and the environmental risk factors in each subtype area in the partitioning mode on the cancer morbidity or mortality risk; the method comprises the following steps of a, b and c:
step a, according to the spatial distribution of the cancer monitoring data in each subtype area, evaluating whether the cancer monitoring data have spatial aggregation in the subtype area by using a formula (2), namely, the closer the monitoring values on the cancer monitoring points are, the more similar the monitoring values are, otherwise, the monitoring value differences on the cancer monitoring points which are farther away are, the formula is as follows:
wherein:
n in formula (2) represents the number of cancer monitoring points in the subtype region, x i And x j Monitoring values at cancer monitoring points i and j, respectively;representing the mean of all monitoring points; spatial weight matrix W ij The method is used for measuring the spatial proximity degree between two cancer monitoring point positions, and the values of elements in a matrix have multiple modes; the most common mode is a 0/1 mode, namely, if the element value in the matrix is 0, the space between the cancer monitoring points i and j is not adjacent, and if the element value is 1, the space between the cancer monitoring points i and j is adjacent; in addition, the spatial weight matrix element values may also be numerical values related to the spatial distance between cancer monitoring points i and j; the larger the value of the spatial correlation index I is, the more obvious spatial aggregation phenomenon exists in the cancer monitoring data in the subtype area;
b, determining the influence of various environmental risk elements in each subtype region on the cancer morbidity or mortality risk in a partition mode through correlation analysis; the higher the relevance index value of a certain environmental risk element is, the greater the influence of the environmental risk element on the cancer morbidity or mortality risk is;
step c, according to the same method as the spatial partition in the region partitioning step, based on various environmental risk factors and cancer monitoring data distribution, determining the partitioning mode of each subtype area, and calculating to obtain the block distribution difference q value of the cancer monitoring data in each subtype area;
step (2), comprehensively considering and determining the aggregation condition, block distribution difference condition and environmental risk factor influence condition of the spatial distribution of the cancer monitoring data in each subtype area or the whole area to be mapped in a partition mode, and forming a mapping model library of different condition combinations of the influence forms of the spatial distribution characteristic of the cancer monitoring data and the environmental risk factor on the cancer morbidity or mortality on the basis of the aggregation condition, the block distribution difference condition and the environmental risk factor influence condition; automatically screening different drawing models suitable for different condition combinations from a drawing model library according to the different condition combinations; if the cancer monitoring data has no obvious space aggregation phenomenon and block distribution difference characteristics but the environmental risk elements have obvious correlation with the cancer monitoring data, selecting a linear regression method or an intelligent algorithm for constructing a relational equation and other models by combining the environmental risk elements; if the cancer monitoring data has a remarkable spatial aggregation phenomenon, selecting a series of models compatible with spatial adjacency relation; if the block distribution of the cancer monitoring data has obvious difference, selecting a series of models compatible with the block attribute of the monitoring data; if the cancer monitoring data has significant spatial clustering phenomenon and significant difference in block distribution, a model combination compatible with spatial adjacency relation and attributes of the block is selected.
Step (3) calculating and determining the accuracy R of the drawing results of different models in each subtype area or the whole area needing drawing in a partition mode by using a k-1 cross validation method 2 ，
X in formula (3) i * Represents the difference between the monitored value at cancer monitoring point i and the mean of all monitoring points, x i ' represents the difference between the monitoring value of the cancer monitoring point i and the calculated value of the cancer monitoring point i obtained by the mapping model, and n represents the number of the cancer monitoring points in the subtype area or the whole area needing mapping; final selection precision value R 2 And the drawing model corresponding to the maximum is the optimal drawing method for the sub-type area or the whole area needing drawing.
The charting model library formed in the step (2) with different condition combinations comprises the following eight conditions: when the environmental risk factors and the cancer monitoring data have obvious correlation, firstly, the cancer monitoring data have no obvious space aggregation phenomenon and block distribution difference characteristics; 2. the cancer monitoring data only has a remarkable spatial aggregation phenomenon but has no block distribution difference characteristics; 3. cancer monitoring data only has block distribution difference characteristics but does not have significant spatial aggregation phenomenon; 4. cancer monitoring data has both significant spatial clustering and significant differences in block distribution; when the environmental risk factors do not have obvious correlation with the cancer monitoring data, the cancer monitoring data do not have obvious space aggregation phenomenon and block distribution difference characteristics; 6. the cancer monitoring data only has a remarkable spatial aggregation phenomenon but has no block distribution difference characteristics; 7. cancer monitoring data only has block distribution difference characteristics but does not have significant spatial aggregation phenomenon; 8. cancer monitoring data has both significant spatial clustering and significant differences in the distribution of the blocks. Significance is a statistical meaning of the results in statistics.
Compared with the prior art, the invention has the advantages that: the invention overcomes the defect that a single method is difficult to accurately fit and estimate the spatial distribution patterns and the spatial distribution rules of the cancer morbidity or mortality of a plurality of regions, comprehensively considers the influence of a plurality of aspects such as the spatial distribution characteristics of the cancer, the influence mechanism of environmental risk factors and the like, integrates a plurality of functions of environmental risk factor selection, partition selection and charting model selection, realizes the fast, intelligent and accurate charting of the cancer in a component form, and successfully realizes the automation of the charting process of the cancer. In addition, the invention uniformly establishes the cancer map making production standard, so that the compilation of the cancer map gradually develops towards the standardization and standardization direction. The production modeling can realize the intelligent expression of the morbidity or mortality of different cancer species in different periods in a plurality of regions.
Drawings
FIG. 1 is a flow chart of a system and method for environmental element-based modular cancer mapping according to the present invention.
Fig. 2 is a graph showing an estimated lung cancer mortality distribution of a woman 2005 in mainland china (except taiwan and south hai islands).
Fig. 3 is a graph showing relative error distribution of lung cancer mortality estimation for women 2005 in mainland china (except taiwan and island of south seas).
Detailed Description
The following takes "modular mapping of lung cancer mortality distribution of women in 2005 in mainland china (except taiwan and south hai island)" as a specific example, and specifically introduces the specific implementation steps of the modular cancer mapping system and method based on environmental elements in the present invention, as shown in fig. 1, the specific implementation process of the modules is as follows:
1. drawing data screening and library building module
Selecting environmental risk factors possibly related to lung cancer death rate of women in 2005 in mainland China (except Taiwan province and Haichoudao islands), from social and economic factors, geographical environmental factors and medical and health factors in mainland China; fusing the female lung cancer death rate monitoring data and the selected possible environmental risk factor data in 2005 in mainland China, and then processing the data of abnormal values, vacancy, negative values and the like of areas with unknown mechanisms, thereby establishing a cancer mapping database; finally determining factors such as national production total value, female education degree, non-agricultural population proportion, female smoking rate, female drinking rate, female meat intake, female vegetable and fruit intake, female overweight rate, elevation, population density, proportion of population over 60 years old, concentration of fine particulate matters in air, proportion of second industry, vegetation coverage and the like as environmental risk factors required for drawing 2005 female lung cancer mortality distribution of Chinese continent by utilizing the correlation between the Chinese continent 2005 female cancer mortality and the environmental risk factors and whether the same spatial distribution pattern exists or not;
2. geographical zone module
Since the pathogenic mechanism of cancer varies from place to place, after environmental risk factors are determined by the mapping data screening and library building module, the region to be mapped is partitioned by the regional partitioning module based on the pathogenic mechanism of cancer and the distribution of cancer data. Based on cancer pathogenesis partitioning, namely starting from environmental risk element distribution related to the lung cancer death risk of women in 2005 in the Chinese continent, partitioning a region to be mapped according to different subtype regions of the environmental risk elements; based on the cancer data distribution partition, the numerical difference is expressed between different subtype areas of the environmental risk elements which are irrelevant to the lung cancer death risk of women in 2005 in mainland China but have more consistent spatial distribution, and the environmental risk element partition mode corresponding to the largest difference is selected to partition the area to be mapped. After a plurality of partition modes are selected, the regional partition module autonomously compares the numerical value space distribution difference of the female lung cancer death rate monitoring data in 2005 in mainland China in different partition modes, and finally determines the urban group partition with the largest difference as the final regional partition mode of the map of the female lung cancer death rate in 2005 in mainland China.
The partitioning steps based on cancer pathogenesis are as follows:
determining distribution of elevation elements most relevant to lung cancer death risk of women in 2005 in mainland China as a standard based on cancer pathogenesis partition through correlation analysis;
and (2) partitioning the elevation data which are continuous numerical data according to the principle that the numerical difference of the cancer monitoring data in the same subtype area is minimum and the difference between different subtype areas is maximum. First setting the number of desired elevational divisions to 5 to 10 zones; then randomly dividing a corresponding number of elevation value intervals as initial partitions, and calculating the difference q between the data values of the female lung cancer death rate monitoring data of 2005 year continental China in subtype areas corresponding to all the value intervals under the initial partitions;
where N represents the number of all cancer monitoring points for the entire mapped region (N = 218), N h Is the number of cancer monitoring points in the subtype h (h =1, \8230;, L) region of a certain environmental risk element, L is the number of divisions of the environmental risk element, σ 2 Is the magnitude of the difference, σ, between all cancer monitor data values for the entire region to be mapped h 2 Is the magnitude of the difference between the values of the cancer monitoring data within subtype h region, q ∈ [0,1]A larger q indicates a larger difference between the cancer monitoring data values in different subtype regions, the better the zoning effect. And then, respectively adding random numbers on the basis of the boundary line value of each numerical value interval of the initial partition to perform partition adjustment, and calculating the change of the q value. And if the q value under the adjusted partition is larger than the q value of the initial partition, taking the adjusted partition as the current optimal partition, and then respectively adding random numbers on the basis of the boundary line value of each numerical value interval of the current optimal partition to perform partition adjustment again. The classification adjustment is circulated for a plurality of times until the adjustment times reach the set iteration times of 300 times. Finally, the optimal partition result of 9 partitions with q value of 0.323 at most is obtained as the elevation.
And (3) the elevation value subsection intervals corresponding to the maximum q values in the step (3) are respectively <31.25 meters, 31.25-48.45 meters, 48.45-53.97 meters, 53.97-292.78 meters, 292.78-325.48 meters, 325.48-1558.15 meters, 1558.15-2063.23 meters, 2063.23-2439.24 meters and >2439.24 meters, and subtype area subsections corresponding to the 9 intervals are used as final subarea modes based on the cancer pathogenesis.
The partitioning step based on the cancer data itself is as follows:
selecting elements of a river basin, a climate zone and an urban group which are irrelevant to the female lung cancer death rate risk in 2005 in mainland China but have consistent spatial distribution, wherein the river basin, the climate zone and the urban group element data are type data and are spatially superposed with female lung cancer death rate monitoring data in 2005 in mainland China one by one, so that the spatial zoning of the cancer monitoring data based on the river basin, the climate zone and the urban group element is realized;
and (2) comparing difference values q between data values of female lung cancer death rate monitoring data in 2005 in mainland China in the sub-type areas of the watershed, the climate zone and the urban group to obtain a final partition mode based on cancer data distribution, wherein the q value of the watershed partition is 0.302, the q value of the climate zone partition is 0.183, the q value of the urban group partition is 0.377, and the urban group partition corresponding to the maximum q value is used as the final partition mode.
The partition pattern with the greatest difference between the cancer monitoring data values in the partition types among the two partition patterns obtained will be the default optimal partition pattern. Comparing the q values of the elevation partition based on the cancer pathogenesis and the urban group partition based on the cancer data distribution, wherein the q value of the urban group partition is the maximum and is 0.377, and therefore the urban group partition is determined to be the default optimal partition mode for drawing the female lung cancer death rate in 2005 in mainland China.
3. Drawing model selection module
The default optimal cancer charting method for each subtype area under the urban grouping partition mode is recommended to be determined by integrating the space distribution characteristics of the female lung cancer death rate monitoring data in 2005 in mainland China and the influence of environmental risk elements determined in the charting data screening and library building module on the female lung cancer death risk in 2005 in continental China; the analysis results of the lung cancer death rate monitoring data spatial distribution characteristics and the environmental risk factor correlation of the women in 2005 in mainland China show that the lung cancer death rate monitoring data of the women in 2005 in mainland China not only show certain aggregation characteristics, but also have larger difference in block distribution, a Kriging method and a multi-level model combined method and a sandwich method which are suitable for the influence form of the spatial distribution characteristics and the environmental risk factors are automatically selected from a model library to respectively map the lung cancer death rate distribution of the women in 2005 in mainland China, the precision evaluation is carried out through cross validation, and finally the distribution mapping of the lung cancer death rate of the women in 2005 in mainland China is carried out by the Kriging method with the highest precision and the multi-level model combined method.
The drawing model selection module is specifically implemented as follows:
after the urban grouping is determined, calculating the influence of the spatial distribution characteristics of the cancer monitoring data and the environmental risk factors in the internal and external subtype areas of the urban grouping on the cancer morbidity or mortality risk in the partitioning mode. The method comprises the following steps:
step a, according to the spatial distribution of the cancer monitoring data in each subtype area, evaluating whether the cancer monitoring data have spatial aggregation in the subtype area by using a formula (2), namely, the closer the monitoring values on the cancer monitoring points are, the more similar the monitoring values are, otherwise, the monitoring value differences on the cancer monitoring points which are farther away are, the formula is as follows:
wherein:
wherein n represents the number of cancer monitoring points in the subtype region, x i And x j The monitored values at cancer monitoring points i and j respectively,mean, spatial weight matrix W representing all monitoring points ij The method is used for measuring the spatial proximity degree between two cancer monitoring point positions, the value of an element in a matrix has a plurality of modes, in the case, a 0/1 mode is used, namely the element value in the matrix is 0 and represents that the space between the cancer monitoring points i and j is not adjacent, and the element value is 1 and represents that the space between the cancer monitoring points i and j is adjacent. Finally, calculating to obtain the city groupThe value of the spatial correlation index I of the internal subtype region is 0.305, and the value of the spatial correlation index I of the subtype region outside the urban group is 0.041, which indicates that the lung cancer death rate monitoring data of women in 2005 in mainland of China has an obvious spatial aggregation phenomenon in the subtype region inside the urban group;
and b, determining the influence of various environmental risk factors in each subtype area in a partition mode on the cancer morbidity or mortality risk through correlation analysis. The greater the value of the relevance index for an environmental risk factor, the greater the impact of that environmental risk factor on the risk of cancer morbidity or mortality.
And c, determining the partition mode of each subtype region based on various environmental risk factors and cancer monitoring data distribution according to the same method as the spatial partition in the regional partition module, and calculating the block distribution difference q value of the female lung cancer death rate monitoring data obtained in 2005 in mainland China to be 0.377.
Step (2), comprehensively considering and determining the aggregation condition, block distribution difference condition and environmental risk factor influence condition of the spatial distribution of the cancer monitoring data in each subtype area or the whole area to be mapped in a partition mode, and forming a mapping model library of different condition combinations of the influence forms of the spatial distribution characteristic of the cancer monitoring data and the environmental risk factor on the cancer morbidity or mortality on the basis of the aggregation condition, the block distribution difference condition and the environmental risk factor influence condition; automatically screening different drawing models suitable for different condition combinations from a drawing model library according to the different condition combinations; if the cancer monitoring data has no obvious space aggregation phenomenon and block distribution difference characteristics but the environmental risk elements have obvious correlation with the cancer monitoring data, selecting a linear regression method or an intelligent algorithm for constructing a relational equation and other models by combining the environmental risk elements; if the cancer monitoring data has a remarkable spatial aggregation phenomenon, selecting a series of models compatible with spatial adjacency relation; if the block distribution of the cancer monitoring data has obvious difference, selecting a series of models compatible with the block attribute of the monitoring data; if the cancer monitoring data has significant spatial clustering phenomenon and significant difference in block distribution, a model combination compatible with spatial adjacency relation and attributes of the block is selected. The female lung cancer death rate monitoring data has obvious correlation with environmental risk factors in 2005 in mainland China, and has obvious space aggregation phenomenon in the interior of the urban group, and has obvious difference between the interior and exterior blocks of the urban group, so the module recommends that the lung cancer death rate of the female in 2005 in mainland China is mapped by using two models of the kriging method and the multilevel model which are suitable for the combination of the environmental risk factors and the cancer monitoring data which have obvious correlation and have obvious space aggregation phenomenon and obvious difference in block distribution, and the cancer monitoring data which have obvious space aggregation phenomenon. The charting model library formed in the step (2) with different condition combinations comprises the following eight conditions: when the environmental risk factors and the cancer monitoring data have obvious correlation, firstly, the cancer monitoring data have no obvious space aggregation phenomenon and block distribution difference characteristics; 2. the cancer monitoring data only has a remarkable spatial aggregation phenomenon but has no block distribution difference characteristics; 3. cancer monitoring data only has block distribution difference characteristics but does not have significant spatial aggregation phenomenon; 4. cancer monitoring data has both significant spatial clustering and significant differences in block distribution; when the environmental risk factors are not significantly correlated with the cancer monitoring data, the cancer monitoring data have no significant space aggregation phenomenon and block distribution difference characteristics; 6. the cancer monitoring data only has a remarkable spatial aggregation phenomenon but has no block distribution difference characteristics; 7. cancer monitoring data only has block distribution difference characteristics but does not have obvious spatial aggregation phenomenon; 8. cancer monitoring data has both significant spatial clustering and significant differences in the distribution of the blocks.
Step (3) calculating and determining the accuracy R of the drawing results of different models in each subtype area or the whole area needing drawing in a partition mode by using a k-1 cross validation method 2 ，
x i * Representative cancer monitoringDifference, x, between the monitored value at point i and the mean of all monitored points i ' represents the difference between the monitoring value of the cancer monitoring point i and the calculated value of the cancer monitoring point i obtained by the mapping model, and n represents the number of the cancer monitoring points in the subtype area or the whole area to be mapped, and is 218. R made by combining kriging method and multi-level model after calculation 2 Value of 0.688, R made by Sandwich method 2 The value is 0.612, and the module finally selects the precision value R 2 And the maximum corresponding kriging method and the drawing model combined with the multi-level model are the optimal drawing method for the whole area needing drawing.
4. Cancer distribution mapping module
After selecting a proper cancer mapping method, the mapping model selection module respectively generates lung cancer death rate distribution data of urban and rural women in 2005 in China continent, generates lung cancer death rate distribution data of the urban and rural women in 2005 in China continent according to population proportion of the urban and rural women, and simultaneously generates mapping error data by using a cross validation method; determining a map layer matching scheme, a legend and an expression mode of a Chinese continent 2005 female lung cancer death rate distribution map, representing administrative units or partition boundary surface feature elements by certain different dotted line symbols or characters, representing the levels of the Chinese continent 2005 female lung cancer death rate and relative errors by adopting chromatic aberration, drawing the Chinese continent 2005 female lung cancer death rate distribution map and an estimation error distribution map, and finally printing the map after quality inspection and draft determination.
The realization process is as follows:
step (1), after a proper cancer mapping method is selected by a mapping model selection module, generating lung cancer death rate distribution data of urban and rural women in 2005 in China continent, then generating lung cancer death rate distribution data of the women in 2005 in China continent according to the urban and rural women population composition proportion to generate the lung cancer death rate distribution data of the women in 2005 in China continent, and simultaneously generating mapping error data by the module by using a cross validation method;
and (2) determining a layer matching scheme, a legend and an expression mode of the cancer distribution map. The graph surface content is regional data and provincial administrative unit boundary data of female lung cancer mortality in 2005 in mainland China, the regional data graph layer of female lung cancer mortality in 2005 in mainland China expressed by 15-level chromatism method is positioned on the first layer surface, provincial administrative unit boundaries expressed in a form of no filling inside the graph spots are positioned on the second layer surface, administrative units are expressed by certain line symbols, and a distribution estimation graph of female lung cancer mortality in 2005 in mainland China is drawn, as shown in fig. 2; an error distribution graph layer of female lung cancer death rate monitoring points of 2005 in mainland China expressed by a 5-level color difference method is positioned on a first level, a Chinese provincial administrative region layer expressed in a form without filling inside a graph spot is positioned on a second level, an administrative unit is expressed by a certain line symbol, and an estimated error distribution graph of female lung cancer death rate distribution of 2005 in mainland China is drawn, as shown in fig. 3; the whole distribution map adopts a Chinese and English marking mode, the graticule is used as a basic control network, and finally, a picture is printed after the quality inspection manuscript is determined.
TABLE 1 color system, legend setup, and font definition for a modular cancer mapping system
Those matters not described in detail in the present specification are well known in the art to which the skilled person pertains.
Claims (8)
1. An intelligent mapping system for cancer based on environmental risk factors, comprising: including charting data screening and build storehouse module, geographical partitioning module, charting model selection module, cancer distribution charting module, wherein:
screening drawing data and building a library module: selecting environmental risk factors related to the morbidity or mortality risk of the cancer to be mapped from social and economic factors, geographic environment factors and medical and health factors of the area to be mapped; fusing the cancer morbidity or mortality monitoring data and the selected environmental risk factor data according to the cancer types, deleting the data of noise, vacancy and logic errors in the data, and establishing a cancer mapping database; finally, determining the environmental risk elements required by the final cancer mapping according to the correlation between the cancer morbidity or mortality monitoring data and the environmental risk element data in the cancer mapping database and whether the various data have the same spatial distribution pattern, namely whether the spatial consistency analysis exists;
a zone partitioning module: for partitioning the area to be mapped based on the cancer pathogenesis and/or the cancer data distribution; the method comprises the following steps of carrying out partitioning based on a cancer pathogenesis mechanism, namely partitioning a region to be mapped according to different subtype regions of environmental risk elements from environmental risk element distribution related to cancer morbidity or mortality; partitioning is carried out based on cancer data distribution, namely, according to the numerical difference of the cancer monitoring data shown between different subtype areas of the environmental risk elements which are irrelevant to the risks of cancer morbidity or mortality and have consistent spatial distribution patterns, the environmental risk element partitioning mode corresponding to the maximum numerical difference of the cancer monitoring data between the subtype areas is selected to partition the area to be mapped; if the user selects a plurality of partition modes, the regional partition module can autonomously compare the numerical difference of the cancer morbidity or mortality monitoring data in different sub-type regions under different partition modes, and determine the partition mode with the largest difference as the final regional partition mode of the cancer mapping; finally, the user obtains a plurality of subtype areas of the area needing drawing in a geographical region module; when the regional division module is selected to be skipped, the drawing system defaults to drawing the region to be drawn under the condition of no division;
a charting model selection module: the mapping module is used for selecting different cancer mapping models for each subtype area in a regionalized mode or the whole area needing mapping under the condition of no partition, and recommending the cancer default optimal mapping models of each subtype area or the whole area needing mapping in the regionalized mode to a user by integrating the spatial distribution characteristics of various cancer monitoring data and the influence of environmental risk factors determined in the mapping data screening and library building module on the cancer morbidity or mortality risk; the model library of the drawing model selection module comprises a plurality of drawing models suitable for different space distribution characteristics and environmental risk factor influence forms;
a cancer distribution mapping module: after the charting model selection module selects a proper cancer charting model, the cancer distribution charting module generates a cancer morbidity/mortality distribution map and a charting error distribution map of a region to be charted according to the selection of a user.
2. A mapping method based on the modular cancer intelligent mapping system of claim 1, characterized in that: the method comprises the following steps:
screening and establishing a database of charting data, namely selecting environmental risk factors possibly related to the morbidity or mortality risk of the cancer to be charted from social and economic factors, geographic environment factors and medical and health factors of a region to be charted; fusing and denoising the cancer morbidity or mortality monitoring data and the selected environmental risk factor data, thereby establishing a cancer mapping database; finally, determining the environmental risk elements required by the final cancer mapping according to the correlation between the monitoring data of the morbidity or the mortality of the cancer and the environmental risk element data in the cancer mapping database and whether the various data have consistent spatial distribution pattern analysis;
a regional division step, namely dividing a region to be mapped based on a cancer pathogenesis and/or cancer data distribution; the method comprises the following steps of carrying out partitioning based on a cancer pathogenesis mechanism, namely partitioning a region to be mapped according to different subtype regions of environmental risk elements from environmental risk element distribution related to cancer morbidity or mortality; partitioning is carried out based on cancer data distribution, namely, according to numerical differences expressed by cancer monitoring data among different subtype zones of the environmental risk elements which are irrelevant to the risks of cancer onset or death but have consistent spatial distribution, the environmental risk element partitioning mode corresponding to the maximum numerical difference is selected to partition the area needing to be mapped; when a plurality of partition modes are selected, the numerical difference of the cancer morbidity or mortality monitoring data in different sub-type regions under different partition modes is automatically compared, and the partition mode with the maximum numerical difference is determined as the final regional partition mode of the cancer mapping;
the method comprises the following steps of (1) selecting a charting model, namely integrating the spatial distribution characteristics of various cancer monitoring data and the influence of environmental risk factors determined in a charting data screening and database building module on the cancer morbidity or mortality risk, and selecting a default optimal cancer charting model for each subtype area under a regional partition mode or the whole area needing charting under the condition of no partition based on a model database comprising a large number of charting models suitable for different spatial distribution characteristics and environmental risk factor influence forms;
and a step of mapping the cancer distribution, namely selecting a proper cancer mapping model, and generating a distribution map of the cancer incidence/rate or death/rate and a mapping error distribution map of a region to be mapped according to the selection of a user.
3. The patterning method according to claim 2, wherein: the determination step of the sub-type region optimal cartographic model comprises the following steps: after the subareas are determined through the geographical division step, the charting model selection step evaluates the influence of the spatial distribution characteristics of the cancer monitoring data and the environmental risk factors on the cancer morbidity or mortality risk in each subtype area; and then, a drawing model selection step automatically selects different drawing models suitable for the space distribution characteristics and the influence form of the environmental risk elements from a model library, respectively draws, evaluates the precision through cross verification, and finally takes the method with the highest precision as the optimal drawing model in the subtype area.
4. The patterning method according to claim 2, wherein: the step of generating the cancer profile comprises: generating the distribution data of the number of diseases/rate or the number of deaths/rate of the cancers to be mapped of different crowds such as city/county, male/female or crowds of different age groups in the area to be mapped, or generating the distribution data of the number of diseases/rate or the number of deaths/rate of the cancers of the whole crowd in the area to be mapped according to the crowd composition ratio by using the cancer distribution data results of different crowds, and simultaneously generating mapping error data by using a cross validation method in the cancer distribution mapping step; determining a map layer matching scheme, a legend and an expression mode of the cancer distribution map, representing administrative units or partition boundary ground object elements by certain different dotted line symbols or characters, representing the cancer morbidity/mortality or the cancer morbidity with relative errors and the grade of the mortality by adopting color difference, drawing the cancer morbidity/mortality or mortality/distribution map and a drawing error distribution map, and finally printing the map after quality inspection.
5. The patterning method according to claim 2, wherein: the step of differentiating based on the pathogenesis of cancer comprises:
determining environmental risk factors related to the risk of cancer onset or death through correlation analysis;
step (2), partitioning each screened environmental risk factor according to the principle that the difference of the numerical values of the cancer monitoring data in the same subtype area is minimum and the difference between different subtype areas is maximum; firstly, setting the number of the expected subareas of the environmental dangerous elements; then randomly dividing the corresponding number of dangerous element value intervals as initial partitions, and calculating the total difference value q between the cancer monitoring data in subtype areas corresponding to all the value intervals under the initial partitions;
wherein N in the formula (1) represents the number of all cancer monitoring points in the whole region to be mapped, N h The number of cancer monitoring points in a subtype region h (h =1, \8230;, L) that is an environmental risk factor, L being the number of divisions of the environmental risk factor, σ 2 Is the magnitude of the numerical difference, σ, between all cancer monitoring data for the entire region to be mapped h 2 Is the magnitude of the numerical difference between cancer monitoring data within subtype h region, q ∈ [0,1]The larger q is, the larger the numerical total difference between the cancer monitoring data in different subtype areas is, and the better the zoning effect is; then, random numbers are added on the basis of the boundary line value of each numerical value interval of the initial partition respectively to carry out partition adjustment, and the change of the q value is calculated; if the q value under the adjusted partition is larger than the q value of the initial partition, the adjusted partition is taken as the current optimal partition, and then the current optimal partition is usedRespectively adding random numbers on the basis of the boundary line value of each numerical value interval of the optimal partition to perform partition adjustment again; the classification adjustment is circulated for a plurality of times until the adjustment times reach the set iteration times; finally, the optimal partition with the maximum q value is used as a partition result of the element;
and (3) comparing the numerical difference values q among the cancer monitoring data in all the selected type subareas of the environmental risk elements, and taking a certain environmental risk element subarea corresponding to the largest q value as a final subarea mode based on the cancer pathogenesis.
6. The patterning method according to claim 2, wherein: the step of partitioning based on the cancer data distribution comprises:
selecting environmental risk elements which are irrelevant to cancer morbidity or mortality and have relatively consistent spatial distribution, and carrying out spatial superposition on the environmental risk elements and the cancer monitoring data one by one, thereby realizing the spatial partitioning of the cancer monitoring data based on each selected environmental risk element;
and (2) comparing the numerical difference values q between the cancer monitoring data in all the subtype areas of the selected environmental risk elements, and taking a certain environmental risk element partition corresponding to the maximum q value as a final partition mode based on the cancer data distribution.
7. The patterning method according to claim 2, wherein: the step of charting model selection further comprises the steps of:
step (1), if the user selects to skip the regional division step, directly evaluating the influence of the spatial distribution characteristics of the cancer monitoring data and environmental risk factors in the whole region to be mapped on the cancer morbidity or mortality risk; otherwise, according to the partitioning result in the regional partitioning step, respectively evaluating the influence of the spatial distribution characteristics of the cancer monitoring data and the environmental risk factors in each subtype area in the partitioning mode on the cancer morbidity or mortality risk; the method comprises the following steps of a, b and c:
step a, according to the spatial distribution of the cancer monitoring data in each subtype area, evaluating whether the cancer monitoring data have spatial aggregation in the subtype area by using a formula (2), namely, the closer the monitoring values on the cancer monitoring points are, the more similar the monitoring values are, otherwise, the monitoring value differences on the cancer monitoring points which are farther away are, the formula is as follows:
wherein:
n in formula (2) represents the number of cancer monitoring points in the subtype region, x i And x j The monitoring values at cancer monitoring points i and j, respectively;representing the mean of all monitoring points; spatial weight matrix W ij The method is used for measuring the spatial proximity degree between two cancer monitoring point positions, and the values of elements in a matrix have multiple modes; the most common mode is a 0/1 mode, namely, if the element value in the matrix is 0, the space between the cancer monitoring points i and j is not adjacent, and if the element value is 1, the space between the cancer monitoring points i and j is adjacent; alternatively, the spatial weight matrix element values may be values related to the spatial distance between cancer monitor points i and j; the larger the value of the spatial correlation index I is, the more obvious spatial aggregation phenomenon exists in the cancer monitoring data in the subtype area;
b, determining the influence of various environmental risk elements in each subtype area on the cancer morbidity or mortality risk in a partition mode through correlation analysis; the higher the relevance index value of a certain environmental risk element is, the greater the influence of the environmental risk element on the cancer morbidity or mortality risk is;
c, determining the partition mode of each subtype area based on various environmental risk factors and the distribution of the cancer monitoring data according to the same method as the spatial partition in the regional partition step, and calculating to obtain the block distribution difference value q of the cancer monitoring data in each subtype area;
step (2), comprehensively considering and determining the aggregation condition, block distribution difference condition and environmental risk factor influence condition of the spatial distribution of the cancer monitoring data in each subtype area or the whole area to be mapped in a partition mode, and forming a mapping model library of different condition combinations of the influence forms of the spatial distribution characteristics of the cancer monitoring data and the environmental risk factors on the cancer morbidity or mortality risk on the basis of the aggregation condition, the block distribution difference condition and the environmental risk factor influence condition; automatically screening different drawing models suitable for different condition combinations from a drawing model library according to the different condition combinations; if the cancer monitoring data has no obvious space aggregation phenomenon and block distribution difference characteristics but the environmental risk elements have obvious correlation with the cancer monitoring data, selecting a linear regression method or an intelligent algorithm for constructing a relational equation and other models by combining the environmental risk elements; if the cancer monitoring data has a remarkable spatial aggregation phenomenon, selecting a series of models compatible with spatial adjacency relation; if the block distribution of the cancer monitoring data has obvious difference, selecting a series of models compatible with the block attribute of the monitoring data; if the cancer monitoring data has a significant spatial aggregation phenomenon and significant difference in block distribution, selecting a model combination compatible with spatial adjacency relation and attributes of the blocks;
step (3) calculating and determining the accuracy R of the drawing results of different models in each subtype area or the whole area needing drawing in a partition mode by using a k-1 cross validation method 2 ，
X in formula (3) i * Represents the difference between the monitored value at cancer monitor point i and the mean of all the monitor points, x i ' represents the difference between the monitoring value of the cancer monitoring point i and the calculated value of the cancer monitoring point i obtained by the mapping model, and n represents the number of the cancer monitoring points in the subtype area or the whole area needing mapping; finally, the product is processedSelecting a precision value R 2 And the drawing model corresponding to the maximum is the optimal drawing method for the sub-type area or the whole area needing drawing.
8. The patterning method according to claim 7, wherein: the charting model library formed in the step (2) with different condition combinations comprises the following eight conditions: when the environmental risk factors and the cancer monitoring data have obvious correlation, firstly, the cancer monitoring data have no obvious space aggregation phenomenon and block distribution difference characteristics; 2. the cancer monitoring data only has a remarkable spatial aggregation phenomenon but has no block distribution difference characteristics; 3. cancer monitoring data only has block distribution difference characteristics but does not have obvious spatial aggregation phenomenon; 4. cancer monitoring data has both significant spatial clustering phenomena and significant differences in block distribution; when the environmental risk factors are not significantly correlated with the cancer monitoring data, the cancer monitoring data have no significant space aggregation phenomenon and block distribution difference characteristics; 6. the cancer monitoring data only has a remarkable spatial aggregation phenomenon but has no block distribution difference characteristics; 7. cancer monitoring data only has block distribution difference characteristics but does not have significant spatial aggregation phenomenon; 8. cancer monitoring data has both significant spatial clustering and significant differences in the distribution of the blocks.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711052882.6A CN107887026B (en) | 2017-11-01 | 2017-11-01 | Component type cancer intelligent mapping system and method based on environmental risk factors |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711052882.6A CN107887026B (en) | 2017-11-01 | 2017-11-01 | Component type cancer intelligent mapping system and method based on environmental risk factors |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107887026A true CN107887026A (en) | 2018-04-06 |
| CN107887026B CN107887026B (en) | 2022-04-05 |
Family
ID=61783323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711052882.6A Active CN107887026B (en) | 2017-11-01 | 2017-11-01 | Component type cancer intelligent mapping system and method based on environmental risk factors |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107887026B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101477533A (en) * | 2008-12-23 | 2009-07-08 | 中国科学院地理科学与资源研究所 | Digital mapping method for uneasily acquiring geographic element spacing gradient information in graticule |
| US20100082362A1 (en) * | 2008-09-17 | 2010-04-01 | Baker Salsbury | Method and Apparatus for Assessing Salient Characteristics of a Community |
| CN104537254A (en) * | 2015-01-07 | 2015-04-22 | 中国科学院地理科学与资源研究所 | Fine drawing method based on social statistical data |
| CN106202883A (en) * | 2016-06-28 | 2016-12-07 | 成都中医药大学 | A kind of method setting up disease cloud atlas based on big data analysis |
-
2017
- 2017-11-01 CN CN201711052882.6A patent/CN107887026B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100082362A1 (en) * | 2008-09-17 | 2010-04-01 | Baker Salsbury | Method and Apparatus for Assessing Salient Characteristics of a Community |
| CN101477533A (en) * | 2008-12-23 | 2009-07-08 | 中国科学院地理科学与资源研究所 | Digital mapping method for uneasily acquiring geographic element spacing gradient information in graticule |
| CN104537254A (en) * | 2015-01-07 | 2015-04-22 | 中国科学院地理科学与资源研究所 | Fine drawing method based on social statistical data |
| CN106202883A (en) * | 2016-06-28 | 2016-12-07 | 成都中医药大学 | A kind of method setting up disease cloud atlas based on big data analysis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107887026B (en) | 2022-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yu et al. | Object-based spatial cluster analysis of urban landscape pattern using nighttime light satellite images: A case study of China | |
| Zahedi et al. | Groundwater quality classification derivation using multi-criteria-decision-making techniques | |
| Abebe | Quantifying urban growth pattern in developing countries using remote sensing and spatial metrics: A case study in Kampala, Uganda | |
| Miller et al. | Digital classification of hillslope position | |
| Chen et al. | Fractal dimensions derived from spatial allometric scaling of urban form | |
| Bielecka | A dasymetric population density map of Poland | |
| Zandbergen et al. | Error propagation models to examine the effects of geocoding quality on spatial analysis of individual-level datasets | |
| Zhang et al. | A global dataset of daily near-surface air temperature at 1-km resolution (2003–2020) | |
| Zhang et al. | Research on the relationship between Urban economic development level and urban spatial structure—A case study of two Chinese cities | |
| Jiang et al. | Box-counting dimension of fractal urban form: stability issues and measurement design | |
| Chen | Fractal systems of central places based on intermittency of space-filling | |
| Pisati | Exploratory spatial data analysis using Stata | |
| Pavía et al. | Can dasymetric mapping significantly improve population data reallocation in a dense urban area? | |
| Haining | Spatial autocorrelation and the quantitative revolution | |
| Pisati | Spatial Data Analysis in Stata an Overview | |
| CN115712634A (en) | City update unit boundary defining method based on spatial entropy and display platform | |
| McGwire et al. | Spatially variable thematic accuracy: Beyond the confusion matrix | |
| CN107944750A (en) | A kind of poverty depth analysis method and system | |
| Tian et al. | Local carbon emission zone construction in the highly urbanized regions: Application of residential and transport CO2 emissions in Shanghai, China | |
| CN113901348A (en) | Oncomelania snail distribution influence factor identification and prediction method based on mathematical model | |
| Ferreira et al. | Methods to calculate urban surface parameters and their relation to the LCZ classification | |
| Maroko et al. | Using geovisualization and geospatial analysis to explore respiratory disease and environmental health justice in New York City | |
| Zhang | Classification of Urban Land Use Based on Graph Theory and Geographic Information System. | |
| CN107887026B (en) | Component type cancer intelligent mapping system and method based on environmental risk factors | |
| Lin et al. | Enhancing Urban Land Use Identification Using Urban Morphology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |