CN108764348A

CN108764348A - Collecting method based on multiple data sources and system

Info

Publication number: CN108764348A
Application number: CN201810539389.5A
Authority: CN
Inventors: 王春龙
Original assignee: Word Of Mouth (beijing) Network Technology Co Ltd
Current assignee: Word Of Mouth (beijing) Network Technology Co Ltd
Priority date: 2018-05-30
Filing date: 2018-05-30
Publication date: 2018-11-06
Anticipated expiration: 2038-05-30
Also published as: CN108764348B

Abstract

The invention discloses a kind of collecting method and system based on multiple data sources, is related to electronic information field, this method includes：The collected multiple data feature values of data characteristics that multiple data sources are directed to acquisition target in this data acquisition are obtained respectively；Multiple data feature values are divided at least one grouping, basis is each grouped the grouping score that corresponding data source determines the grouping for the feature weight value of acquisition target respectively；According to the grouping score of each grouping, determine collection result of the data characteristics of acquisition target in this data acquisition, and the feature weight value to each data source for acquisition target is modified, to use each data source to be directed to the revised feature weight value of acquisition target in data acquisition next time.According to this method, collection result of the corresponding data source of each grouping for the feature weight value of acquisition target and the data characteristics of acquisition target in data acquisition can be more accurately obtained.

Description

Collecting method based on multiple data sources and system

Technical field

The present invention relates to electronic information fields, and in particular to a kind of collecting method based on multiple data sources and is System.

Background technology

With the development of science and technology, target identification technology combined of multi-sensor information is widely used in military, calculating The fields such as machine vision, remote sensing images, the meaning with extensive practical application.The target identification technology energy of feature based grade fusion Enough in characteristic information various complicated, uncertain, that fusion multiple sensors provide, reduce data volume bandwidth, raising The precision and efficiency of identification.Wherein, statistical nature is one of statistical basic conception, overall being studied with mathematical statistics method When, people are of concern actually and the overall each individual of non-constitutive itself.Statistical nature has quantative attribute and attributive character Point, wherein quantative attribute is divided into metrology features and count feature, quantative attribute that can directly be indicated with numerical value again, example Such as, plant height of the size dimension of element, wheat etc. is metrology features.

Currently used feature level fusing method is roughly divided into five major class：Probability theory statistical method, logic reasoning, god Through network method, the fusion method of feature based extraction and based on the fusion method of search.Probability theory system is simply introduced below Several algorithms in meter method.Wherein, (1) weighted mean method：Assuming that have n sensor, after same process feature extraction, Characteristic value after the weighted average of the characteristic value x of certain dimension is merged as the dimension, wherein weight can regard not simultaneous interpretation as The measurement of sensor accuracy.(2) Bayes' assessment：Bayes (Bayes) method of estimation is for multi-sensor information fusion When, it is the various uncertain informations that multisensor provides to be expressed as probability, and utilize Bayes condition probability formulas in probability theory A kind of method that they are handled.Known to prior probability, bayesian criterion is best fusion criterion, can Provide accurate fusion results.(3) Kalman filtering method is one " optimizing autoregression data processing method ", Kalman filtering Suitable for the dynamic low level redundant sensor information of real time fusion.The statistical property of this method measurement model is carried out recursion and is gone out Determine the optimal estimation fusion data under statistical significance.If system has linear kinetic model, and system noise and biography Sensor noise is all the White Noise Model of Gaussian Profile, and the method optimal is estimated for what data fusion provided under unique solvent meaning The recursion characteristic of meter, Kalman makes system processing not need a large amount of data storage and calculate.

But inventor has found in the implementation of the present invention：It, must thing although weighted mean method simple, intuitive Detailed analysis first is carried out to each sensor, obtains its weight.And the accuracy of each sensor is tieed up in different characteristic It is all different on degree, so the acquisition of weight becomes Major Difficulties.In addition, about Bayes' assessment, in practical application In, each sensor is difficult to obtain required prior probability, and this point greatly limits the application of bayesian criterion.About karr Graceful filter method, for many nonlinear systems, there are no a set of stringent Filtering Formulas.It can be seen that lacking in the prior art It is a kind of simple and practical, and can accurately obtain the feature that multiple sensors are directed to some acquisition target in data acquisition The collecting method of characteristic value after weighted value and collected multiple data feature values fusions.

Invention content

In view of the above problems, it is proposed that the present invention overcoming the above problem in order to provide one kind or solves at least partly State a kind of collecting method and system based on multiple data sources of problem.

According to an aspect of the invention, there is provided a kind of collecting method based on multiple data sources, including：Respectively Obtain the collected multiple data characteristicses of data characteristics that multiple data sources are directed to acquisition target in this data acquisition Value；Multiple data feature values are divided at least one grouping, basis is each grouped corresponding data source for acquisition pair respectively The feature weight value of elephant determines the grouping score of the grouping；According to the grouping score of each grouping, the data of acquisition target are determined Collection result of the feature in this data acquisition, and the feature weight value to each data source for acquisition target carries out It corrects, to use each data source to be directed to the revised feature weight of acquisition target in data acquisition next time Value.

Optionally, wherein described the step of the multiple data feature values are divided at least one grouping specifically includes：

Clustering processing is carried out to the multiple data feature values according to default clustering rule, according to clustering processing result by institute It states multiple data feature values and is divided at least one grouping.

Optionally, wherein the basis respectively is each grouped corresponding data source and is weighed for the feature of the acquisition target Weight values determine that the step of grouping score of the grouping specifically includes：

For each grouping, determine that the corresponding each data source of the grouping is directed to each feature of the acquisition target respectively Weighted value, each feature weight value according to default operation rule to each data source for the acquisition target are transported It calculates, the grouping score of the grouping is obtained according to operation result.

Optionally, wherein the step that each data source is modified for the feature weight value of the acquisition target Suddenly it specifically includes：

If the grouping score of the corresponding grouping of data source is higher than the first predetermined threshold value, increases the data source and adopted for described Collect the feature weight value of object；

If the grouping score of the corresponding grouping of data source is less than the second predetermined threshold value, reduces the data source and adopted for described Collect the feature weight value of object.

Optionally, wherein described to obtain multiple data sources respectively in this data acquisition for acquisition target The step of data characteristics collected multiple data feature values, specifically includes：Multiple data sources are obtained respectively in (n+1)th data The collected multiple data feature values of data characteristics of acquisition target are directed in gatherer process；Wherein, n is just whole not less than 0 Number；

And the basis respectively is each grouped corresponding data source and is somebody's turn to do for the feature weight value determination of the acquisition target The step of grouping score of grouping, specifically includes：Basis is each grouped corresponding data source in n-th data acquisition respectively In for the feature weight value of the data characteristics of the acquisition target determine the grouping score of the grouping；

Then described the step of being modified for the feature weight value of the acquisition target to each data source, specifically includes： Determine each data source in (n+1)th data acquisition for the data characteristics of the acquisition target according to following formula Feature weight value；

Wn+1=(Wn+m)/(n+1)；Wherein, Wn+1 is that the data source acquires described in (n+1)th data acquisition The feature weight value of the data characteristics of object, Wn are that the data source is directed to the acquisition target in n-th data acquisition Data characteristics feature weight value, the value of m determines according to the grouping score of the corresponding grouping of the data source.

Optionally, wherein before the method executes, further comprise：

The initial characteristics weighted value that each data source is directed to the acquisition target is pre-set, by the initial characteristics weight It is worth the feature weight value for the data characteristics for being directed to the acquisition target in the 0th data acquisition as data source.

Optionally, wherein when the data characteristics of the acquisition target includes multiple dimensions, each data source is for described The feature weight value of acquisition target includes：The feature weight value of multiple different dimensions for corresponding respectively to the acquisition target；

Then the data characteristics for obtaining multiple data sources respectively for acquisition target in this data acquisition is adopted The step of multiple data feature values collected, specifically includes：Multiple data sources are obtained respectively to be directed in this data acquisition The collected multiple data feature values of data characteristics of the default dimension of acquisition target；

The basis respectively is each grouped corresponding data source and determines this point for the feature weight value of the acquisition target The step of grouping score of group, specifically includes：Basis is each grouped corresponding data source for the default of the acquisition target respectively The feature weight value of dimension determines the grouping score of the grouping；

Then collection result of the data characteristics of the determination acquisition target in this data acquisition, and to each The step of a data source is modified for the feature weight value of the acquisition target specifically includes：

Determine collection result of the data characteristics of the default dimension of the acquisition target in this data acquisition, and Feature weight value to each data source for the default dimension of the acquisition target is modified.

Optionally, wherein the acquisition target includes：Temperature, humidity, heartbeat, and/or entity；The data source includes： Camera sensing device, and/or voice operated sensor.

According to another aspect of the present invention, a kind of collecting method based on multiple data sources is provided, including：Point Multiple shops's characteristic values that multiple data sources arrive in this data acquisition for shops's collection apparatus of shops are not obtained；

It is at least one grouping by the multiple shops's feature value division, basis is each grouped corresponding data source needle respectively The grouping score of the grouping is determined to the feature weight value of the shops；

According to the grouping score of each grouping, shops's feature adopting in this data acquisition of the shops is determined Collection as a result, and each data source is modified for the feature weight value of the shops, to be acquired in data next time The revised feature weight value of the shops is directed in journey using each data source.

Optionally, wherein described to specifically include the multiple shops's feature value division for the step of at least one grouping：

Clustering processing is carried out to the multiple shops's characteristic value according to default clustering rule, according to clustering processing result by institute It is at least one grouping to state multiple shops's feature value divisions.

Optionally, wherein the basis respectively is each grouped the feature weight value that corresponding data source is directed to the shops The step of grouping score for determining the grouping, specifically includes：

For each grouping, determine that the corresponding each data source of the grouping is directed to each feature weight of the shops respectively Value, each feature weight value according to default operation rule to each data source for the shops carry out operation, according to Operation result obtains the grouping score of the grouping.

Optionally, wherein described the step of being modified for the feature weight value of the shops to each data source has Body includes：

If the grouping score of the corresponding grouping of data source is higher than the first predetermined threshold value, increases the data source and be directed to the door The feature weight value in shop；

If the grouping score of the corresponding grouping of data source is less than the second predetermined threshold value, reduces the data source and be directed to the door The feature weight value in shop.

Optionally, wherein the shops for obtaining multiple data sources respectively and being directed to shops in this data acquisition Collection apparatus to multiple shops's characteristic value acquisition target shops feature shops features the step of specifically include：It obtains respectively multiple Multiple shops's characteristic values that data source arrives in (n+1)th data acquisition for shops's collection apparatus of shops；Wherein, n For the positive integer not less than 0；

And the basis respectively is each grouped corresponding data source and determines the grouping for the feature weight value of the shops Grouping score the step of specifically include：Basis is each grouped corresponding data source needle in n-th data acquisition respectively The grouping score of the grouping is determined to the feature weight value of shops's feature of the shops；

Then described the step of being modified for the feature weight value of the shops to each data source, specifically includes：According to Following formula determines each data source in (n+1)th data acquisition for the feature weight of shops's feature of the shops Value；

Wn+1=(Wn+m)/(n+1)；Wherein, Wn+1 is described in the data source is directed in (n+1)th data acquisition The feature weight value of shops's feature of shops, Wn are the door that the data source is directed to the shops in n-th data acquisition The value of the feature weight value of shop feature, m is determined according to the grouping score of the corresponding grouping of the data source.

Optionally, wherein described to obtain multiple data sources respectively when shops's feature of the shops includes multiple dimensions In this data acquisition for shops shops's collection apparatus to multiple shops's characteristic values the step of specifically include：Point Do not obtain multiple data sources in this data acquisition for shops acquisition target shops's collection apparatus arrive it is multiple Shops's characteristic value；

The basis respectively is each grouped corresponding data source and determines the grouping for the feature weight value of the shops The step of being grouped score specifically includes：Respectively according to the spy for being each grouped the dimension that corresponding data source is directed in the shops Sign weighted value determines the grouping score of the grouping；

Then collection result of the shops's feature of the determination shops in this data acquisition, and to each number The step of being modified for the feature weight value of the shops according to source specifically includes：

Determine collection result of the shops's feature of the dimension in the shops in this data acquisition, and to each The feature weight value for the dimension that a data source is directed in the shops is modified.

Optionally, wherein shops's feature of the shops includes following at least one dimension：Shops's type dimension, shops In commodity dimension and/or electronic ticket dimension；

The data source includes：The data source of data source, the data source that user provides and APP offers that businessman provides.

According to another aspect of the present invention, a kind of data collecting system based on multiple data sources is provided, including：First Acquisition module acquires in this data acquisition for the data characteristics of acquisition target suitable for obtaining multiple data sources respectively The multiple data feature values arrived；

First determining module, suitable for the multiple data feature values are divided at least one grouping, respectively according to each It is grouped the grouping score that corresponding data source determines the grouping for the feature weight value of the acquisition target；

First correcting module is suitable for the grouping score according to each grouping, determines that the data characteristics of the acquisition target exists Collection result in this data acquisition, and the feature weight value to each data source for the acquisition target is repaiied Just, so as in data acquisition next time use each data source be directed to the acquisition target revised feature weight Value.

Optionally, wherein first determining module is particularly adapted to：

Optionally, wherein first correcting module is particularly adapted to：

Optionally, wherein first acquisition module is particularly adapted to：Multiple data sources are obtained respectively in (n+1)th data The collected multiple data feature values of data characteristics of acquisition target are directed in gatherer process；Wherein, n is just whole not less than 0 Number；

And first determining module is particularly adapted to：Basis is each grouped corresponding data source and is adopted in n-th data respectively Feature weight value during collection for the data characteristics of the acquisition target determines the grouping score of the grouping；

Then first correcting module is particularly adapted to：Determine that each data source is adopted in (n+1)th data according to following formula For the feature weight value of the data characteristics of the acquisition target during collection；

Wn+1=(Wn+m)/(n+1)；Wherein, Wn+1 is described in the data source is directed in (n+1)th data acquisition The feature weight value of the data characteristics of acquisition target, Wn are that the data source is directed to the acquisition in n-th data acquisition The value of the feature weight value of the data characteristics of object, m is determined according to the grouping score of the corresponding grouping of the data source.

Optionally, wherein described device further comprises setup module, is suitable for：

Then first acquisition module is particularly adapted to：Multiple data sources are obtained respectively to be directed in this data acquisition The collected multiple data feature values of data characteristics of the default dimension of acquisition target；

First determining module is particularly adapted to：Basis is each grouped corresponding data source and is directed to the acquisition target respectively The feature weight value of default dimension determine the grouping score of the grouping；

Then first correcting module is particularly adapted to：

According to another aspect of the present invention, a kind of data collecting system based on multiple data sources is provided, including：

Second acquisition module, suitable for obtaining multiple data sources respectively in this data acquisition for the shops of shops Multiple shops's characteristic values that collection apparatus arrives；

Second determining module is suitable for the multiple shops's feature value division being at least one grouping, respectively according to each It is grouped the grouping score that corresponding data source determines the grouping for the feature weight value of the shops；

Second correcting module is suitable for the grouping score according to each grouping, determines shops's feature of the shops at this Collection result in data acquisition, and the feature weight value to each data source for the shops is modified, so as to The revised feature weight value of the shops is directed to using each data source in data acquisition next time.

Optionally, wherein second determining module is particularly adapted to：

Optionally, wherein second correcting module is particularly adapted to：

Optionally, wherein second acquisition module is particularly adapted to：Multiple data sources are obtained respectively in (n+1)th data The multiple shops's characteristic values arrived for shops's collection apparatus of shops in gatherer process；Wherein, n is the positive integer not less than 0；

And second determining module is particularly adapted to：Basis is each grouped corresponding data source and is adopted in n-th data respectively Feature weight value during collection for shops's feature of the shops determines the grouping score of the grouping；

Then second correcting module is particularly adapted to：Determine that each data source is adopted in (n+1)th data according to following formula For the feature weight value of shops's feature of the shops during collection；

Optionally, wherein when shops's feature of the shops includes multiple dimensions, the second acquisition module is particularly adapted to： Obtain respectively multiple data sources in this data acquisition for shops acquisition target shops's collection apparatus arrive it is more A shops's characteristic value；

Second determining module is particularly adapted to：Basis is each grouped corresponding data source and is directed to the shops respectively In the feature weight value of the dimension determine the grouping score of the grouping；

Then second correcting module is particularly adapted to：

Another aspect according to the present invention, provides a kind of electronic equipment, including：Processor, memory, communication interface and Communication bus, the processor, the memory and the communication interface complete mutual communication by the communication bus；

For the memory for storing an at least executable instruction, it is as above that the executable instruction makes the processor execute The corresponding operation of the collecting method based on multiple data sources stated.

Another aspect according to the present invention provides another electronic equipment, including：Processor, memory, communication interface And communication bus, the processor, the memory and the communication interface complete mutual lead to by the communication bus Letter；

According to the present invention in another aspect, provide a kind of computer storage media, be stored in the storage medium to A few executable instruction, the executable instruction make processor execute such as the above-mentioned collecting method based on multiple data sources Corresponding operation.

According to the present invention in another aspect, provide another computer storage media, be stored in the storage medium An at least executable instruction, the executable instruction make processor execute such as the above-mentioned data acquisition side based on multiple data sources The corresponding operation of method.

According to collecting method and system provided by the invention based on multiple data sources, by obtaining multiple data sources It is directed to the collected multiple data feature values of data characteristics of acquisition target in this data acquisition, and is advised according to cluster Multiple data feature values are then divided at least one grouping, basis is each grouped corresponding data source respectively for above-mentioned grouping The grouping score of the grouping is determined for the feature weight value of acquisition target, is adopted to which the corresponding data source of each grouping to be directed to Feature weight value and the grouping score of the grouping for collecting object are interrelated, finally according to the grouping score of each grouping, Determine collection result of the data characteristics of acquisition target in this data acquisition, and to each data source for acquisition pair The feature weight value of elephant is modified, to use each data source for acquisition target in data acquisition next time Revised feature weight value.It, can be by the corresponding data source of each grouping of constantly iteration for acquisition according to this method The feature weight value of object to more accurately obtain features described above weighted value, and determines most according to voting mechanism Whole data acquisition results, rather than the method simply by calculating average value, so as to according to above-mentioned accurate feature Weighted value more accurately determines collection result of the data characteristics of acquisition target in this data acquisition, this method letter Single practical, the collection result accuracy finally obtained is higher.

Above description is only the general introduction of technical solution of the present invention, in order to better understand the technical means of the present invention, And can be implemented in accordance with the contents of the specification, and in order to allow above and other objects of the present invention, feature and advantage can It is clearer and more comprehensible, below the special specific implementation mode for lifting the present invention.

Description of the drawings

By reading the detailed description of hereafter preferred embodiment, various other advantages and benefit are common for this field Technical staff will become clear.Attached drawing only for the purpose of illustrating preferred embodiments, and is not considered as to the present invention Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings：

Fig. 1 shows a kind of flow for collecting method based on multiple data sources that the embodiment of the present invention one provides Figure；

Fig. 2 shows one flow chart elements of a kind of collecting method based on multiple data sources provided in an embodiment of the present invention Frame figure；

Fig. 3 shows another of another collecting method based on multiple data sources provided in an embodiment of the present invention Flow frame diagram；

Fig. 4 shows a kind of flow of collecting method based on multiple data sources provided by Embodiment 2 of the present invention Figure；

Fig. 5 shows a kind of structure for data collecting system based on multiple data sources that the embodiment of the present invention three provides Figure；

Fig. 6 shows the structure for another data collecting system based on multiple data sources that the embodiment of the present invention four provides Figure；

Fig. 7 shows the structural schematic diagram for a kind of electronic equipment that the embodiment of the present invention five provides；

Fig. 8 shows the structural schematic diagram for another electronic equipment that the embodiment of the present invention six provides.

Specific implementation mode

The exemplary embodiment of the disclosure is more fully described below with reference to accompanying drawings.Although showing the disclosure in attached drawing Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here It is limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure Completely it is communicated to those skilled in the art.

Embodiment one

Fig. 1 shows a kind of flow for collecting method based on multiple data sources that the embodiment of the present invention one provides Figure.As shown in Figure 1, this method includes：

Step S110：The initial characteristics weighted value that each data source is directed to acquisition target is pre-set, initial characteristics are weighed Weight values are as data source for the feature weight value of the data characteristics of acquisition target in the 0th data acquisition.

Wherein, which could be provided as impartial numerical value, for example can be set as the numerical value such as 0.5,1,2 Or it is set as other specific numerical value.It, can also be according to the parameter value of each data source and sensitive other than aforesaid way Property flexibly sets the initial characteristics weighted value, can neatly select this those skilled in the art, refuses herein With limitation.

Wherein, this step is an optional step can also omit this step in other embodiments.For example, when each A data source itself then can be omitted this step when having had initial characteristics weighted value.

Step S120：The data characteristics that multiple data sources are directed to acquisition target in this data acquisition is obtained respectively Collected multiple data feature values.

Wherein, above-mentioned data source includes but not limited at least one of the following：Camera sensing device, and/or acoustic control pass Sensor.The sensor refers to a kind of detection device, can experience measured information, and can be by the information experienced by certain Rule is for conversion into electric signal or the information output form needed for other, to meet the transmission of information, processing, storage, display, note The requirements such as record and control.The characteristics of sensor includes：Micromation, digitlization, intelligence, multifunction, systematization, networking. It is to realize the primary link for detecting and automatically controlling automatically.Wherein, during above-mentioned acquisition target includes but not limited to below extremely It is one few：Temperature, humidity, heartbeat, and/or entity.Such as when acquisition target be entity in commodity, for some of the commodity Characteristic dimension can obtain the valence that multiple data sources are directed to commodity in this gatherer process respectively such as price dimension Multiple price values that lattice collection apparatus arrives.In order to more clearly be fully described technical scheme of the present invention, Fig. 2 shows this hairs A kind of one flow frame diagram of collecting method based on multiple data sources that bright embodiment provides.As shown in Fig. 2, its In, it is assumed that above-mentioned multiple data sources are multiple sensors, respectively sensor i, sensor j, and sensor k is then directed to and a certain adopts The data characteristics for collecting object can obtain the data feature values of above three sensor acquisition respectively.

Optionally, when the data characteristics of above-mentioned acquisition target includes multiple dimensions, then multiple data can be obtained respectively Source is in this data acquisition for the collected multiple data feature values of data characteristics of the default dimension of acquisition target. Wherein, which can be any of above-mentioned multiple dimensions.As shown in Fig. 2, the data characteristics of such as acquisition target Including 4 dimensions, that is, it is characterized 1, feature 2, feature 3, feature 4, then can obtain multiple data sources respectively and be adopted in this data For the collected multiple data of data characteristics of the default dimension of any of above-mentioned 4 dimensions of acquisition target during collection Characteristic value.It is for another example directed to the data characteristics of commodity, when the data characteristics of commodity includes that price, weight, usage time etc. are multiple Dimension can obtain multiple data sources for the data characteristics of any of the above-described dimension and be adopted in this data respectively in this step For the collected multiple data feature values of data characteristics of the dimension of acquisition target during collection.

Step S130：Multiple data feature values are divided at least one grouping, basis is each grouped corresponding number respectively The grouping score of the grouping is determined for the feature weight value of acquisition target according to source.

It specifically, can be according to preset clustering rule to multiple data when being grouped to multiple data feature values Characteristic value carries out clustering processing, and multiple data feature values are then divided at least one grouping according to clustering processing result.On To state clustering rule such as can be clustering rule based on numberical range, based on the how many clustering rule of number of words or can also be Clustering rule based on type.For example the data feature values that data area can be not much different are divided into same group, or will Data feature values with identical number of words are divided into same group, or similar type, the data feature values of identical type are divided It is one group, in short, the present invention does not limit the specific type of clustering rule, can will has by preset clustering rule similar The data feature values of attributive character or similar size are divided into the same grouping, to indirectly respectively according to each grouping Corresponding data source determines the grouping score of the grouping for the feature weight value of acquisition target.Fig. 3 shows implementation of the present invention The flow frame diagram for another collecting method based on multiple data sources that example provides.As shown in figure 3, sensor i and biography Sensor j is abc for the collected data feature values of data characteristics of acquisition target in this data acquisition, sensing The collected data feature values of device k are abd, so as to which above-mentioned two data feature values are divided into one group, and by sensor k Collected data feature values abd is divided into another group.As shown in Figure 2 and Figure 3, the operation being grouped to data feature values, can To be completed by each Fusion Features device in data fusion device.

Wherein, the grouping score being each grouped can each be grouped corresponding data source with basis respectively and be directed to acquisition target The relationship that is positively correlated of feature weight value, i.e. the corresponding data source of the grouping gets over for the feature weight value of acquisition target The score of height, the grouping is higher.Specifically, basis is each grouped feature power of the corresponding data source for acquisition target for the first time When weight values determine the grouping score of the grouping, it can determine the grouping score being each grouped according to initial characteristics weighted value, connect Being grouped for the grouping can be determined when getting off to calculate grouping score every time according to last revised feature weight value Point, wherein the concrete mode being modified for the feature weight value of acquisition target to each data source will be in following step S140 Middle detailed description, details are not described herein.Specifically, when calculating grouping score, it can be directed to each grouping, determine this point respectively The corresponding each data source of group is directed to each feature weight value of above-mentioned acquisition target, then according to default operation rule to each Data source carries out operation for each feature weight value of acquisition target, and obtains being grouped for the grouping according to operation result Point.Wherein, above-mentioned operation rule can be additivity rule, operation rule of averaging, linear regression computation rule etc., specifically Ground, the operation rule can determine according to the mode clustered in above-mentioned clustering rule, can also be according to specific business scenario It determines, those skilled in the art can flexibly select, the present embodiment is not limited herein.

Optionally, when the data characteristics of acquisition target includes multiple dimensions, each data source is for the acquisition target Feature weight value includes：The feature weight value of multiple different dimensions for corresponding respectively to the acquisition target.Such as shown in Fig. 2, When the data characteristics of acquisition target includes 4 dimensions, for sensor i, it is directed to the feature weight value packet of the acquisition target The feature weight value for 4 different dimensions for corresponding respectively to the acquisition target is included, that is, is directed to feature 1, feature 2, feature 3, feature 4 Feature weight value be respectively：W_(i,1)、W_(i,2)、W_(i,3)、W_(i,4), the acquisition pair is corresponded respectively to about other sensors in Fig. 2 The feature weight value of the different dimensions of elephant is referred to content shown in Fig. 2, and details are not described herein.Therefore, when acquisition target When data characteristics includes multiple dimensions, basis it each can be grouped the default dimension that corresponding data source is directed to acquisition target respectively Feature weight value determine the grouping score of the grouping.The default dimension can be any one in above-mentioned multiple dimensions.By In each data source during gathered data, for the data characteristics of different dimensions, the sensitivity of gathered data and Accuracy can occur to change accordingly, therefore when the data characteristics of acquisition target includes multiple dimensions, by using root respectively The grouping score of the grouping is determined for the feature weight value of the default dimension of acquisition target according to the corresponding data source of each grouping Mode can improve the accuracy of determining score, to obtain more accurate data acquisition results.

By implementing the step, the corresponding data source of each grouping for the feature weight value of acquisition target and can be somebody's turn to do The grouping score of grouping is interrelated, and the feature that corresponding data source is directed to acquisition target is each grouped so as to basis Weighted value determines collection result of the data characteristics of acquisition target in this data acquisition.

Step S140：According to the grouping score of each grouping, determine that the data characteristics of acquisition target is acquired in this data Collection result in the process, and the feature weight value to each data source for acquisition target is modified, so as to next time The revised feature weight value of acquisition target is directed in data acquisition using each data source.

Optionally, when each data source includes for the feature weight value of acquisition target：It is multiple to correspond respectively to acquisition pair The feature weight value of the different dimensions of elephant；It can then determine the data characteristics of the default dimension of acquisition target at this in this step Collection result in secondary data acquisition, and to each data source for acquisition target default dimension feature weight value into Row is corrected.

Specifically, such as in ballot activity, when there is multiple alternative objects, according to voting mechanism, due to obtaining poll The confidence level highest of highest object, so the highest object of poll would generally be selected to obtain as selected object.Similarly, exist In each grouping obtained, for the grouping corresponding to the larger feature weight value of numerical value, corresponding each data source is directed to The collected data feature values confidence level of data characteristics of acquisition target is higher, so as to choose the corresponding data source of the grouping It is acquired in this data as the data characteristics of acquisition target for the collected data feature values of data characteristics of acquisition target Collection result in the process.According to above-mentioned principle, the grouping that can choose the grouping highest scoring of acquisition in this step corresponds to Data source for acquisition target data characteristics data feature values, the data characteristics as acquisition target adopts in this data Collection result during collection.As shown in figure 3, feature s is directed to, since sensor i, sensor j are directed to the data of acquisition target The grouping score being grouped where the data feature values that collection apparatus arrives is more than sensor k and is acquired for the data characteristics of acquisition target The grouping score being grouped where the data feature values arrived, it is possible to which data feature values abc is determined as to the data of acquisition target Collection result of the feature in this data acquisition.

Further, when being modified for the feature weight value of acquisition target to each data source, if data source pair The grouping score for the grouping answered is higher than the first predetermined threshold value, shows the corresponding data source of the grouping during gathered data, needle It is high to the data feature values accuracy of the data characteristics acquisition of acquisition target, then increase the feature that the data source is directed to acquisition target Weighted value.Correspondingly, if the grouping score of the corresponding grouping of data source is less than the second predetermined threshold value, show that the grouping is corresponding For data source during gathered data, the data feature values accuracy for the data characteristics acquisition of acquisition target is low, then reduces The data source is directed to the feature weight value of acquisition target, and the feature weight of acquisition target is directed to by increasing or decreasing data source Value, so as to complete the amendment to each data source for the feature weight value of acquisition target.Wherein, the above-mentioned first default threshold Value can be the same or different with the second predetermined threshold value, and those skilled in the art can be according to specific service conditions come specific Setting.

Specifically, above by increase or decrease data source be directed to acquisition target feature weight value, to complete To each data source for acquisition target feature weight value it is modified during, can also be according to specific formula to each Data source is more accurately corrected for the feature weight value of acquisition target, to realize to each data source for acquisition The feature weight value of object constantly iteration with can be according to being each grouped corresponding number in data acquisition each time Last time revised feature weight value according to source for acquisition target determines the grouping score being each grouped.It next will be to upper Formula is stated to be described in detail：Assuming that in the step s 120, obtaining multiple data sources respectively in (n+1)th data acquisition In be directed to acquisition target the collected multiple data feature values of data characteristics；Wherein, n is the positive integer not less than 0；Then in step In rapid S130, basis corresponding data source each can be grouped in n-th data acquisition for acquisition target respectively The feature weight value of data characteristics determines the grouping score of the grouping；The then above-mentioned feature that acquisition target is directed to each data source Weighted value can be with when being modified：According to formula W_n+1=(W_n+ m)/(n+1) determine that each data source is adopted in (n+1)th data For the feature weight value of the data characteristics of acquisition target during collection；Wherein, W_n+1It is adopted in (n+1)th data for the data source For the feature weight value of the data characteristics of acquisition target, W during collection_nIt is the data source in n-th data acquisition For the feature weight value of the data characteristics of acquisition target, the value of m is true according to the grouping score of the corresponding grouping of the data source It is fixed.Specifically, m can be positive number, negative, the numerical value such as 0, and the size of concrete numerical value can be grouped according to the data source is corresponding Divide and determined with the comparison result of the first predetermined threshold value and the second predetermined threshold value, if above-mentioned grouping score is higher than the first default threshold Value, then m could be provided as positive value (such as+1), and m values could be provided as the difference with grouping score and the first predetermined threshold value Directly proportional value；If above-mentioned grouping score is less than the second predetermined threshold value, m could be provided as the numerical value such as negative value (such as 0 or -1), Similarly, the absolute value of m could be provided as the value directly proportional to grouping score and the difference of the second predetermined threshold value.According to above-mentioned public affairs Formula W_n+1=(W_n+ m)/(n+1), it can determine that each data source is directed to the number of acquisition target in (n+1)th data acquisition According to the feature weight value of feature, to complete to be directed to the number of acquisition target in n-th data acquisition to each data source According to the amendment of the feature weight value of feature.As shown in figure 3, feature s is directed to, according to above-mentioned formula to each data source for acquisition The operating process that the feature weight value of object is modified can be completed by data source characteristic weight calculator.Such as Data source i can pass through formula W for the feature weight value for acquisition target in data acquisition '_(i,1)= (W_(i,1)+ m)/(n+1) be modified, wherein W '_(i,1)For (i.e. (n+1)th time) the revised number in this data acquisition The feature weight value of feature 1, W are directed to according to source i_(i,1)For in upper primary data acquisition revised data source i for spy The feature weight value of sign 1.For the feature for acquisition target of other data source j, data source k in data acquisition Weighted value can also be modified according to the formula in such as Fig. 3, and details are not described herein.

According to above-mentioned formula W_n+1=(W_n+ m)/(n+1), and after every data acquisition for completing a data source, each number The weight of the data characteristics of acquisition target can all be occurred dynamically to change according to source, by being directed to acquisition target to the data source Data feature values constantly iteration, to make the weighted value ad infinitum level off to theoretical weighted value, and then make the number next time It is more accurate for the data acquisition results of the data characteristics of acquisition target according to source.

According to the collecting method provided in this embodiment based on multiple data sources, by presetting each data source first For the initial characteristics weighted value of acquisition target, multiple data sources are then obtained in this data acquisition for acquisition pair The collected multiple data feature values of data characteristics of elephant, and multiple data feature values are divided at least one according to clustering rule A grouping, for above-mentioned grouping respectively according to be each grouped that feature weight value of the corresponding data source for acquisition target determine should The grouping score of grouping, to which the corresponding data source of each grouping is directed to the feature weight value of acquisition target and point of the grouping Group score is interrelated, finally according to the grouping score of each grouping, determines the data characteristics of acquisition target in this number Be modified for the feature weight value of acquisition target according to the collection result in gatherer process, and to each data source, so as to It is directed to the revised feature weight value of acquisition target in data acquisition using each data source next time.According to the party Method can be directed to the feature weight value of acquisition target, to more accurate by the corresponding data source of each grouping of constantly iteration True acquisition features described above weighted value, and determine final data acquisition results according to voting mechanism, rather than it is simple Method of the ground by calculating average value, so as to more accurately determine acquisition target according to above-mentioned accurate feature weight value Collection result of the data characteristics in this data acquisition, this method is simple and practical, and the collection result finally obtained is accurate Exactness is higher.

Embodiment two

Fig. 4 shows a kind of flow of collecting method based on multiple data sources provided by Embodiment 2 of the present invention Figure.As shown in figure 4, this method includes：

Step S410 obtains shops's collection apparatus that multiple data sources are directed to shops in this data acquisition respectively The multiple shops's characteristic values arrived.

Wherein, above-mentioned data source includes but not limited at least one of the following：The data source of businessman's offer, user provide Data source and APP provide data source.Wherein, the data source that businessman provides such as can be shops's letter that businessman fills in Breath, the data source that user provides such as can be the shops's information filled in from client user oneself, the data source that APP is provided For example can be shops's information that APP small two is filled according to client shops.In addition, shops's feature of above-mentioned shops includes but unlimited At least one dimension in following：Commodity dimension in shops's type dimension, shops and/or electronic ticket dimension.Such as door Shop type dimension can obtain shops collection apparatus of multiple data sources for shops in this data acquisition and arrive respectively Multiple shops's characteristic values, for example can be food and drink class shops, clothing shops, furniture shops etc..

Further, it when shops's feature of above-mentioned shops includes multiple dimensions, can obtain respectively in this step more Multiple shops's features that a data source arrives in this data acquisition for shops's collection apparatus of the acquisition target of shops Value.Wherein, above-mentioned acquisition target can be shops's feature of any of the above-described dimension.Such as shops's type feature comprising When multiple dimensions such as shops's name, shops address, shops's longitude and latitude, multiple data sources can be obtained respectively and are acquired in this data It is directed to multiple shops's characteristic values that shops's collection apparatus of any of the above-described dimension arrives in the process.

Multiple shops's feature value divisions are at least one grouping by step S420, and basis is each grouped corresponding number respectively The grouping score of the grouping is determined for the feature weight value of shops according to source.

When being clustered to shops, clustering processing can be carried out to multiple shops's characteristic values according to default clustering rule, according to Multiple shops's feature value divisions are at least one grouping by clustering processing result.Wherein, above-mentioned clustering rule can such as be base In the clustering rule of numberical range, based on the how many clustering rule of number of words or can also be the clustering rule based on type.Than It is such as directed to the name feature of shops, in this data acquisition, the data source that businessman provides is referred to as data source 1 herein, Its this feature value obtained is yellow braised chicken；The data source that user provides, herein referred to as data source 2, this feature value obtained For little Huang braised chicken；The data source that APP is provided is referred to as data source 3 herein, and this feature value obtained is yellow braised chicken.Due to data Source 1, the characteristic value that data source 3 obtains is yellow braised chicken, then the spy that can be obtained data source 1, data source 3 according to clustering rule Value indicative " yellow braised chicken " is divided into one group, and the characteristic value " little Huang braised chicken " that data source 2 obtains is divided into another group.For another example, for The address feature of shops, in this data acquisition, the characteristic value that data source 1 obtains is " Xihu District of Hangzhou City ", data The characteristic value that the characteristic value that source 2 obtains is " Xihu District of Hangzhou City colleges and universities road ", data source 3 obtains also is that " Xihu District of Hangzhou City is high Data 2, the characteristic value " Xihu District of Hangzhou City colleges and universities road " that data 3 obtain then can be divided into one by school road " according to clustering rule The characteristic value " Xihu District of Hangzhou City " that data source 1 obtains is divided into another group by group.It, can also about other shops's features It is at least one grouping by multiple shops's feature value divisions according to above-mentioned clustering rule, no longer an one kind is stated herein.

After by multiple shops's feature value divisions at least one grouping, basis is each grouped corresponding data source respectively The grouping score of the grouping is determined for the feature weight value of shops.Wherein, the grouping score being each grouped can be with each point The relationship that the corresponding data source of group is positively correlated for the feature weight value of shops, the i.e. corresponding data source of the grouping are directed to door The feature weight value in shop is higher, and the score of the grouping is higher.Specifically, it is directed to for the first time according to being each grouped corresponding data source When the feature weight value of shops's feature determines the grouping score of the grouping, each point can be determined according to initial characteristics weighted value The grouping score of group, which can artificially be set, can be with when next calculating grouping score every time According to the grouping score of the grouping is determined to last revised feature weight value, wherein to each data source for acquisition The concrete mode that the feature weight value of object is modified will be described in detail in following step S230, and details are not described herein.Tool Body, when calculating grouping score, it can be directed to each grouping, it is above-mentioned to determine that the corresponding each data source of the grouping is directed to respectively Then each feature weight value of shops's feature is directed to each data source according to default operation rule each spy of shops's feature It levies weighted value and carries out operation, the grouping score of the grouping is obtained according to operation result.Wherein, above-mentioned operation rule can be to be added Rule, operation rule of averaging, linear regression computation rule etc., specifically, the operation rule can be advised according to above-mentioned cluster The mode clustered in then determines, can also be determined according to specific business scenario in shops, and those skilled in the art can be with Flexibly selection, the present embodiment are not limited herein.

Optionally, when shops's feature of above-mentioned shops includes multiple dimensions, since each data source is in gathered data In the process, for the data characteristics of different dimensions, the sensitivity and accuracy of gathered data can occur to change accordingly, because This can determine the grouping according to the feature weight value for being each grouped the dimension that corresponding data source is directed in shops respectively It is grouped score, to improve the accuracy of determining score, and then obtains shops's feature of more accurate shops in this number According to the collection result in gatherer process.Such as when shops's feature includes multiple dimensions such as shops's title, shops address, shops's longitude and latitude When spending, can respectively according to be each grouped that feature weight value of the corresponding data source for above three dimension in shops determine should The grouping score of grouping, being determined respectively according to the data value for being each grouped each dimension that corresponding data source is directed in shops should The grouping score process of grouping can carry out independently of each other, without interfering with each other.

Step S430 determines shops's feature of shops in this data acquisition according to the grouping score of each grouping In collection result, and each data source is modified for the feature weight value of shops, to be acquired in data next time The revised feature weight value of shops is directed to using each data source in the process.

Wherein, when shops's feature of above-mentioned shops includes multiple dimensions, it may be determined that the shops of the dimension in shops Collection result of the feature in this data acquisition, and to each data source for the feature weight of the dimension in shops Value is modified, to improve modified accuracy.

Specifically, in this data acquisition, due to it is each grouping obtain shops shops's characteristic value it is credible Degree increases with the increase of grouping score, according to above-mentioned principle so as to corresponding more according to the grouping of grouping highest scoring A data source is directed to the collected multiple shops's characteristic values of shops's characteristic value of shops in this data acquisition, to determine Collection result of the shops's feature of shops in this data acquisition.Such as the characteristic value that data source 1, data source 3 obtain The grouping score organized where the characteristic value " little Huang braised chicken " that the score organized where " yellow braised chicken " is obtained higher than data source 2, then can be with It is " yellow braised chicken " in collection result of the name feature for determining shops in this data acquisition.

After collection result of the shops's feature for determining shops in this data acquisition, to each data source needle The feature weight value of shops is modified, to use each data source for shops in data acquisition next time Revised feature weight value.Specifically, if the grouping score of the corresponding grouping of data source is higher than the first predetermined threshold value, increase The data source is directed to the feature weight value of shops；If the grouping score of the corresponding grouping of data source is less than the second predetermined threshold value, Reduce the feature weight value that the data source is directed to shops.Above-mentioned first predetermined threshold value, the second predetermined threshold value could be provided as equal Value, may be set to be unequal value, the size of concrete numerical value can be according to actual by those skilled in the art Service conditions is neatly configured.

Specifically, above by increase or decrease data source be directed to shops feature weight value, so as to complete When being directed to the amendment of the feature weight value of shops to each data source, each data source can also be directed to according to specific formula The feature weight value of shops is more accurately corrected, and the feature weight value of shops is directed to not to each data source to realize Disconnected ground iteration is with can be according to be each grouped that corresponding data source is directed to shops upper one in data acquisition each time Secondary revised feature weight value determines the grouping score being each grouped.Next above-mentioned formula will be described in detail, Assuming that in step S210, the data characteristics that multiple data sources are directed to shops in (n+1)th data acquisition is obtained respectively Collected multiple data feature values；Wherein, n is the positive integer not less than 0；It, can be respectively according to every then in step S220 A feature weight value for being grouped corresponding data source for the data characteristics of shops in n-th data acquisition determines this point The grouping score of group；It can be according to following formula when then the above-mentioned feature weight value for being directed to shops to each data source is modified W_n+1=(W_n+ m)/(n+1) determine each data source in (n+1)th data acquisition be directed to shops feature weight value；Its In, W_n+1The feature weight value of shops, W are directed in (n+1)th data acquisition for the data source_nIt is the data source n-th The feature weight value of shops is directed in secondary data acquisition, the value of m can be according to the grouping of the corresponding grouping of the data source Score determines that specifically, m can be positive number, negative, the numerical value such as 0, and the size of concrete numerical value can be corresponding according to the data source Score and the comparison result of the first predetermined threshold value and the second predetermined threshold value are grouped to determine, if above-mentioned grouping score is higher than first Predetermined threshold value, then m could be provided as positive value (such as+1), and m values could be provided as and grouping score and the first predetermined threshold value The directly proportional value of difference；If above-mentioned grouping score is less than the second predetermined threshold value, m could be provided as the numerical value such as negative value (such as 0 Or -1), similarly, the absolute value of m could be provided as the value directly proportional to grouping score and the difference of the second predetermined threshold value.According to Above-mentioned formula W_n+1=(W_n+ m)/(n+1), it can determine each data source in (n+1)th data acquisition for shops Feature weight value, to complete repairing for the feature weight value for being directed to shops in n-th data acquisition to each data source Just.According to above-mentioned formula, after every data acquisition for completing a data source, data source can all send out the weighted value of shops Raw dynamic variation, by the data feature values constantly iteration to the data source for shops, to keep the weighted value unlimited Ground levels off to the theory weighted value, so make the data source next time for the data characteristics of shops data acquisition results more It is accurate to add.

According to the collecting method provided in this embodiment based on multiple data sources, by obtaining multiple data sources respectively For multiple shops's characteristic values for arriving of shops's collection apparatus of shops in this data acquisition, and by above-mentioned multiple shops Feature value division is at least one grouping, and basis is each grouped corresponding data source and is determined for the feature weight value of shops respectively The grouping score of the grouping, to according to the grouping score of each grouping, determine that shops's feature of shops is acquired in this data Collection result in the process, and the feature weight value to each data source for shops is modified, so as in data next time The revised feature weight value of shops is directed in gatherer process using each data source.It can be to each grouping according to this method Corresponding data source is iterated for the feature weight value of shops, to obtain more accurate feature weight value, and root Determine final data acquisition results according to voting mechanism, rather than the method simply by calculating average value, so as to It is enough more accurately to determine shops's feature of shops in this data acquisition according to above-mentioned accurate feature weight value Collection result, this method is simple and practical, and the collection result accuracy finally obtained is higher.

Embodiment three

Fig. 5 shows a kind of structure for data collecting system based on multiple data sources that the embodiment of the present invention three provides Figure, the system include：

First acquisition module 52 is directed to acquisition target suitable for obtaining multiple data sources respectively in this data acquisition The collected multiple data feature values of data characteristics；

First determining module 53, suitable for the multiple data feature values are divided at least one grouping, respectively according to every A grouping score for being grouped corresponding data source and determining the grouping for the feature weight value of the acquisition target；

First correcting module 54 is suitable for the grouping score according to each grouping, determines the data characteristics of the acquisition target Collection result in this data acquisition, and the feature weight value to each data source for the acquisition target carries out It corrects, to use each data source to be weighed for the revised feature of the acquisition target in data acquisition next time Weight values.

Optionally, wherein first determining module 53 is particularly adapted to：

Optionally, wherein first correcting module 54 is particularly adapted to：

Optionally, wherein first acquisition module 52 is particularly adapted to：Multiple data sources are obtained respectively in (n+1)th number According to the collected multiple data feature values of data characteristics for being directed to acquisition target in gatherer process；Wherein, n is just not less than 0 Integer；

And first determining module 53 is particularly adapted to：Basis is each grouped corresponding data source in n-th data respectively Feature weight value in gatherer process for the data characteristics of the acquisition target determines the grouping score of the grouping；

Then first correcting module 54 is particularly adapted to：Determine each data source in (n+1)th data according to following formula For the feature weight value of the data characteristics of the acquisition target in gatherer process；

W_n+1=(W_n+m)/(n+1)；Wherein, W_n+1It is adopted for described in (n+1)th data acquisition for the data source Collect the feature weight value of the data characteristics of object, W_nFor the data source for the acquisition pair in n-th data acquisition The value of the feature weight value of the data characteristics of elephant, m is determined according to the grouping score of the corresponding grouping of the data source.

Optionally, wherein described device further comprises setup module 51, is suitable for：

Then first acquisition module 52 is particularly adapted to：Multiple data sources needle in this data acquisition is obtained respectively To the collected multiple data feature values of the data characteristics of the default dimension of acquisition target；

First determining module 53 is particularly adapted to：Basis is each grouped corresponding data source for the acquisition pair respectively The feature weight value of the default dimension of elephant determines the grouping score of the grouping；

Then first correcting module 54 is particularly adapted to：

Example IV

Fig. 6 shows the structure for another data collecting system based on multiple data sources that the embodiment of the present invention four provides Figure, the system include：

Second acquisition module 61, suitable for obtaining multiple data sources respectively in this data acquisition for the door of shops Multiple shops's characteristic values that shop collection apparatus arrives；

Second determining module 62 is suitable for the multiple shops's feature value division being at least one grouping, respectively according to every A grouping score for being grouped corresponding data source and determining the grouping for the feature weight value of the shops；

Second correcting module 63 is suitable for the grouping score according to each grouping, determines shops's feature of the shops at this Collection result in secondary data acquisition, and the feature weight value to each data source for the shops is modified, with Just the revised feature weight value of the shops is directed to using each data source in data acquisition next time.

Optionally, wherein second determining module 62 is particularly adapted to：

Optionally, wherein second correcting module 63 is particularly adapted to：

Optionally, wherein second acquisition module 61 is particularly adapted to：Multiple data sources are obtained respectively in (n+1)th number The multiple shops's characteristic values arrived according to the shops's collection apparatus for being directed to shops in gatherer process；Wherein, n is the positive integer not less than 0；

And second determining module 62 is particularly adapted to：Basis is each grouped corresponding data source in n-th data respectively Feature weight value in gatherer process for shops's feature of the shops determines the grouping score of the grouping；

Then second correcting module 63 is particularly adapted to：Determine each data source in (n+1)th data according to following formula For the feature weight value of shops's feature of the shops in gatherer process；

W_n+1=(W_n+m)/(n+1)；Wherein, W_n+1The door is directed in (n+1)th data acquisition for the data source The feature weight value of shops's feature in shop, W_nThe shops of the shops is directed in n-th data acquisition for the data source The value of the feature weight value of feature, m is determined according to the grouping score of the corresponding grouping of the data source.

Optionally, wherein when shops's feature of the shops includes multiple dimensions, the second acquisition module 61 is specifically suitable In：Obtain what multiple data sources arrived in this data acquisition for shops's collection apparatus of the acquisition target of shops respectively Multiple shops's characteristic values；

Second determining module 62 is particularly adapted to：Basis is each grouped corresponding data source and is directed to the door respectively The feature weight value of the dimension in shop determines the grouping score of the grouping；

Then second correcting module 63 is particularly adapted to：

Embodiment five

Fig. 7 shows the structural schematic diagram of according to embodiments of the present invention five a kind of electronic equipment, present invention specific implementation Example does not limit the specific implementation of electronic equipment.

As shown in fig. 7, the electronic equipment may include：Processor (processor) 702, communication interface (Communications Interface) 706, memory (memory) 704 and communication bus 708.

Wherein：

Processor 702, communication interface 706 and memory 704 complete mutual communication by communication bus 708.

Communication interface 706, for being communicated with the network element of miscellaneous equipment such as client or other servers etc..

Processor 702 can specifically execute the above-mentioned data acquisition side based on multiple data sources for executing program 710 Correlation step in method embodiment.

Specifically, program 710 may include program code, which includes computer-managed instruction.

Processor 702 may be central processor CPU or specific integrated circuit ASIC (Application Specific Integrated Circuit), or be arranged to implement the integrated electricity of one or more of the embodiment of the present invention Road.The one or more processors that electronic equipment includes can be same type of processor, such as one or more CPU；Also may be used To be different types of processor, such as one or more CPU and one or more ASIC.

Memory 704, for storing program 710.Memory 704 may include high-speed RAM memory, it is also possible to further include Nonvolatile memory (non-volatile memory), for example, at least a magnetic disk storage.

Program 710 specifically can be used for so that processor 702 executes following operation：

It is collected that data characteristics of multiple data sources for acquisition target in this data acquisition is obtained respectively Multiple data feature values；

The multiple data feature values are divided at least one grouping, basis is each grouped corresponding data source needle respectively The grouping score of the grouping is determined to the feature weight value of the acquisition target；

According to the grouping score of each grouping, determine the data characteristics of the acquisition target in this data acquisition Collection result, and each data source is modified for the feature weight value of the acquisition target, so as in next number According to the revised feature weight value for using each data source to be directed to the acquisition target in gatherer process.

Embodiment six

Fig. 8 shows the structural schematic diagram for another electronic equipment that the embodiment of the present invention six provides.The present invention is specifically real Example is applied not limit the specific implementation of electronic equipment.

As shown in figure 8, the electronic equipment may include：Processor (processor) 802, communication interface (Communications Interface) 806, memory (memory) 804 and communication bus 808.

Wherein：

Processor 802, communication interface 806 and memory 804 complete mutual communication by communication bus 808.

Communication interface 806, for being communicated with the network element of miscellaneous equipment such as client or other servers etc..

Processor 802 can specifically execute the above-mentioned data acquisition side based on multiple data sources for executing program 810 Correlation step in method embodiment.

Specifically, program 810 may include program code, which includes computer-managed instruction.

Processor 802 may be central processor CPU or specific integrated circuit ASIC (Application Specific Integrated Circuit), or be arranged to implement the integrated electricity of one or more of the embodiment of the present invention Road.The one or more processors that electronic equipment includes can be same type of processor, such as one or more CPU；Also may be used To be different types of processor, such as one or more CPU and one or more ASIC.

Memory 804, for storing program 810.Memory 804 may include high-speed RAM memory, it is also possible to further include Nonvolatile memory (non-volatile memory), for example, at least a magnetic disk storage.

Program 810 specifically can be used for so that processor 802 executes following operation：

Obtain respectively multiple data sources in this data acquisition for shops shops's collection apparatus arrive it is multiple Shops's characteristic value；

Algorithm and display be not inherently related to any certain computer, virtual system or miscellaneous equipment provided herein. Various general-purpose systems can also be used together with teaching based on this.As described above, it constructs required by this kind of system Structure be obvious.In addition, the present invention is not also directed to any certain programmed language.It should be understood that can utilize various Programming language realizes the content of invention described herein, and the description done above to language-specific is to disclose this hair Bright preferred forms.

In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that the implementation of the present invention Example can be put into practice without these specific details.In some instances, well known method, structure is not been shown in detail And technology, so as not to obscure the understanding of this description.

Similarly, it should be understood that in order to simplify the disclosure and help to understand one or more of each inventive aspect, Above in the description of exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes In example, figure or descriptions thereof.However, the method for the disclosure should be construed to reflect following intention：It is i.e. required to protect Shield the present invention claims the more features of feature than being expressly recited in each claim.More precisely, as following Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore, Thus the claims for following specific implementation mode are expressly incorporated in the specific implementation mode, wherein each claim itself All as a separate embodiment of the present invention.

Those skilled in the art, which are appreciated that, to carry out adaptively the module in the equipment in embodiment Change and they are arranged in the one or more equipment different from the embodiment.It can be the module or list in embodiment Member or component be combined into a module or unit or component, and can be divided into addition multiple submodule or subelement or Sub-component.Other than such feature and/or at least some of process or unit exclude each other, it may be used any Combination is disclosed to all features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so to appoint Where all processes or unit of method or equipment are combined.Unless expressly stated otherwise, this specification (including adjoint power Profit requires, abstract and attached drawing) disclosed in each feature can be by providing the alternative features of identical, equivalent or similar purpose come generation It replaces.

In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments In included certain features rather than other feature, but the combination of the feature of different embodiments means in of the invention Within the scope of and form different embodiments.For example, in the following claims, embodiment claimed is appointed One of meaning mode can use in any combination.

The all parts embodiment of the present invention can be with hardware realization, or to run on one or more processors Software module realize, or realized with combination thereof.It will be understood by those of skill in the art that can use in practice Microprocessor or digital signal processor (DSP) realize the prize drawing according to the ... of the embodiment of the present invention based on voice input information The some or all functions of some or all components in system.The present invention is also implemented as being retouched here for executing The some or all equipment or program of device (for example, computer program and computer program product) for the method stated. It is such to realize that the program of the present invention may be stored on the computer-readable medium, or can have one or more signal Form.Such signal can be downloaded from internet website and be obtained, either provide on carrier signal or with it is any its He provides form.

It should be noted that the present invention will be described rather than limits the invention for above-described embodiment, and ability Field technique personnel can design alternative embodiment without departing from the scope of the appended claims.In the claims, Any reference mark between bracket should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not Element or step listed in the claims.Word "a" or "an" before element does not exclude the presence of multiple such Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real It is existing.In the unit claims listing several devices, several in these devices can be by the same hardware branch To embody.The use of word first, second, and third does not indicate that any sequence.These words can be explained and be run after fame Claim.

Claims

1. a kind of collecting method based on multiple data sources, including：

It is collected multiple that data characteristics of multiple data sources for acquisition target in this data acquisition is obtained respectively Data feature values；

The multiple data feature values are divided at least one grouping, basis is each grouped corresponding data source and is directed to institute respectively The feature weight value for stating acquisition target determines the grouping score of the grouping；

According to the grouping score of each grouping, data characteristics the adopting in this data acquisition of the acquisition target is determined Collection as a result, and each data source is modified for the feature weight value of the acquisition target, to be adopted in data next time The revised feature weight value of the acquisition target is directed to during collection using each data source.

2. described that the multiple data feature values are divided at least one grouping according to the method described in claim 1, wherein The step of specifically include：

Clustering processing is carried out to the multiple data feature values according to default clustering rule, it will be described more according to clustering processing result A data feature value division is at least one grouping.

3. method according to claim 1 or 2, wherein the basis respectively is each grouped corresponding data source and is directed to institute It states the step of the feature weight value of acquisition target determines the grouping score of the grouping and specifically includes：

For each grouping, determine that the corresponding each data source of the grouping is directed to each feature weight of the acquisition target respectively Value, each feature weight value according to default operation rule to each data source for the acquisition target carry out operation, The grouping score of the grouping is obtained according to operation result.

4. a kind of collecting method based on multiple data sources, including：

Multiple shops that multiple data sources arrive in this data acquisition for shops's collection apparatus of shops are obtained respectively Characteristic value；

It is at least one grouping by the multiple shops's feature value division, basis is each grouped corresponding data source and is directed to institute respectively The feature weight value for stating shops determines the grouping score of the grouping；

According to the grouping score of each grouping, acquisition knot of the shops's feature of the shops in this data acquisition is determined Fruit, and the feature weight value to each data source for the shops is modified, so as in data acquisition next time The revised feature weight value of the shops is directed to using each data source.

5. a kind of data collecting system based on multiple data sources, including：

First acquisition module, suitable for obtaining multiple data sources respectively in this data acquisition for the data of acquisition target Multiple data feature values that collection apparatus arrives；

First determining module, suitable for the multiple data feature values are divided at least one grouping, respectively according to each grouping Corresponding data source determines the grouping score of the grouping for the feature weight value of the acquisition target；

First correcting module is suitable for the grouping score according to each grouping, determines the data characteristics of the acquisition target at this Collection result in data acquisition, and the feature weight value to each data source for the acquisition target is modified, To use each data source to be directed to the revised feature weight value of the acquisition target in data acquisition next time.

6. a kind of data collecting system based on multiple data sources, including：

Second acquisition module, suitable for obtaining multiple data sources respectively in this data acquisition for shops's feature of shops Collected multiple shops's characteristic values；

Second determining module is suitable for the multiple shops's feature value division being at least one grouping, respectively according to each grouping Corresponding data source determines the grouping score of the grouping for the feature weight value of the shops；

Second correcting module is suitable for the grouping score according to each grouping, determines shops's feature of the shops in this data Collection result in gatherer process, and the feature weight value to each data source for the shops is modified, so as under The revised feature weight value of the shops is directed in data acquisition using each data source.

7. a kind of electronic equipment, including：Processor, memory, communication interface and communication bus, the processor, the storage Device and the communication interface complete mutual communication by the communication bus；

The memory makes the processor execute as right is wanted for storing an at least executable instruction, the executable instruction Ask the corresponding operation of the collecting method based on multiple data sources described in any one of 1-3.

8. a kind of electronic equipment, including：Processor, memory, communication interface and communication bus, the processor, the storage Device and the communication interface complete mutual communication by the communication bus；

The memory makes the processor execute as right is wanted for storing an at least executable instruction, the executable instruction Ask the corresponding operation of the collecting method based on multiple data sources described in 4.

9. a kind of computer storage media, an at least executable instruction, the executable instruction are stored in the storage medium Processor is set to execute the corresponding behaviour of the collecting method as claimed in any one of claims 1-3 based on multiple data sources Make.

10. a kind of computer storage media, an at least executable instruction, the executable instruction are stored in the storage medium Processor is set to execute the corresponding operation of collecting method as claimed in claim 4 based on multiple data sources.