CN103167505A - A cell data channel configuration method and system - Google Patents

Abstract

The invention discloses a cell data channel configuration method and system. The method includes the steps of: collecting cell service data; classifying cells with the same service characteristics into one category according to the collected cell service data; adopting a regression method for each type of cell For a cell, an association relationship between channel and parameter configurations and network indicators is established; according to the association relationship between channel and parameter configurations and network indicators for each type of cell, resource allocation is performed for the corresponding cell. The system includes: a data acquisition module, a community classification module and a configuration module. By adopting the technical solution of the present invention, the data service channel can be reasonably planned, which not only guarantees the quality of the data service, improves user perception, but also improves the utilization efficiency of the data channel as much as possible, and improves the quality of the voice service.

Description

A cell data channel configuration method and system

technical field

The invention relates to the field of wireless technology, in particular to a cell data channel configuration method and system.

Background technique

The Campbell algorithm is a data traffic channel configuration method widely used at present. The key is to introduce service resource intensity. Different services occupy different wireless resources. High-speed services occupy more resources, while low-speed services occupy less resources. The Campbell model defines service resource intensity to reflect the occupation of wireless resources by different services. In the GPRS (General Packet Radio Service, general packet radio service technology) system, the service resource intensity is the demand for bandwidth, with the capacity bandwidth of a single time slot being 1, the service resource intensity of each service is assumed to be A _i , and the service resource intensity of each service is The traffic volume is E _i , then the service resource intensity and traffic volume of the virtual service (Campbell service) are

${\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{\underset{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; ,</span>$ ${\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; =</span>\frac{{<span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\underset{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}$

于提供了个坎贝尔信道给坎贝尔业务使用，对n_x按照Erlang B公式(如2％的呼损)计算的业务量除以每用户的坎贝尔业务量，即是小区可以支持的用户数。If a cell can provide n physical time slots for GPRS services, it is quite

In order to provide a Campbell channel for the Campbell service, divide the service volume calculated according to the Erlang B formula (such as 2% call loss) by the Campbell service volume of each user for n _x , that is, the number of users that the cell can support.

Similar to this, Chinese patent CN102083215A also proposes a method for data channel configuration, including: obtaining the current data volume of a cell that needs to be configured with PDCH (Packet Data Channel, packet data channel); and according to the obtained current data volume Determine the number of PDCH configurations currently required by the cell with the set quality of service condition; the set quality of service condition represents the data transmission rate requirement of the cell; and according to the determined current required PDCH configuration number, the cell Perform PDCH configuration.

The traditional wireless channel resource planning uses the Campbell algorithm, which has many artificial assumptions, such as: each service request data is continuously transmitted with the average bandwidth, and the number of channels requested by the terminal each time is the minimum number of channels that can meet the bandwidth requirements wait. The actual data service is multiplexed at the Block level of the RLC (Radio Link Control, radio link control) layer, which is bursty. The first assumption reduces channel multiplexing and usage efficiency, while the second assumption reduces hub efficiency, because half of the terminal application time slots are applied for according to the hardware capability level, and services are not differentiated.

Due to the differences in service allocation and data service user behavior characteristics in different communities, the service characteristics and configuration scheme obtained from the data of a certain community can achieve high accuracy for this community, but for other communities, due to the different service characteristics The difference greatly reduces the accuracy. In practical applications, there are problems such as the impossibility of large-scale rapid deployment.

The method mentioned in the patent CN102083215A is similar to the Campbell method, but it has been improved in this respect considering the service quality condition. The problem is also that the configuration scheme is targeted at a certain community and cannot be promoted on a large scale.

Contents of the invention

The purpose of the present invention is to provide a cell data channel configuration method and system, which can reasonably plan the data service channel, not only guarantee the quality of the data service, improve user perception, but also improve the utilization efficiency of the data channel as much as possible, improve The quality of the voice service.

An embodiment of the present invention provides a cell data channel configuration method including steps:

Collect community service data; according to the collected community service data, classify the communities with the same service characteristics into one category; use the regression method for each type of community to establish the correlation between channel and parameter configuration and network indicators; The channel and parameter configuration of each type of cell are associated with network indicators, and resource configuration is performed on the corresponding cell.

The cell data channel configuration method of the present invention also includes the following steps:

Evaluate the change of key performance indicators before and after the adjustment of the resource configuration of the corresponding cell, if the evaluation result is not satisfactory, re-establish the correlation between the channel and parameter configuration and the network index until the evaluation result is satisfactory.

The cell data channel configuration method of the present invention acquires the KPI of the resource configuration of the cell, and judges whether the KPI reaches a preset value, and if so, the evaluation result is ideal, otherwise, the evaluation result is unsatisfactory.

In the cell data channel configuration method of the present invention, cell service data is collected through a data snapshot system.

In the cell data channel configuration method of the present invention, the service characteristics include service data flow, cell service frequency and cell service duration.

The cell data channel configuration method of the present invention adopts a data mining method and a clustering algorithm to classify cells with the same service characteristics into one category.

The cell data channel configuration method of the present invention further includes: preprocessing the collected cell service data to obtain the processed cell service data; using the processed cell service data as input, first using the clustering algorithm Two step to obtain the target The number of categories, using the number of target categories as input, using the clustering algorithm K-means to obtain the cell type number; verify the cell type number, if the verification result is not satisfactory, adjust the parameters of Two step and/or K-means until the result ideal.

In the cell data channel configuration method of the present invention, the step of establishing the relationship between the channel and parameter configuration and the network index further includes: establishing a prediction model for the network index TRF239; establishing a prediction model for the configuration parameter CDEF; calculating the corresponding CDED parameter configuration.

In the cell data channel configuration method of the present invention, the step of establishing a prediction model for the network index TRF239 further includes, selecting the predictive independent variable; performing data conversion on the independent variable; taking the data converted by the independent variable as input, and using multiple linear regression The algorithm is used to calculate the result; analyze the result, if the evaluation index cannot be reached, re-select the independent variable, or perform other data conversion on the independent variable, or choose other fitting algorithms, until the evaluation index is reached.

In the cell data channel configuration method of the present invention, the step of establishing a prediction model for the configuration parameter CDEF further includes, using TRF_239, the configuration parameters of the lifting domain network, the cell voice single time slot load and the data single time slot load as independent variables, and using linear regression Algorithm to calculate CDEF.

The cell data channel configuration method of the present invention further includes: obtaining the cell type mapping rule set by adopting a decision tree method according to the cell service data.

The embodiment of the present invention also provides a cell data channel configuration system, including: a data collection module, a cell classification module, and a configuration module; a data collection module, for collecting cell service data; a cell classification module, for The collected cell service data classifies the cells with the same service characteristics into one category; the configuration module is used to establish the correlation between the corresponding channel and parameter configuration and network indicators for each type of cell divided by the cell classification module .

The cell data channel configuration system of the present invention also includes an evaluation module, which is used to evaluate the resource configuration adjustment of the cell after the resource configuration is performed on the corresponding cell according to the channel and parameter configuration for each type of cell and the association relationship between the parameter configuration and the network index. If the changes of key performance indicators before and after are unsatisfactory, call the configuration module to re-establish the correlation between channel and parameter configuration and network indicators, until the evaluation result of the change of key performance indicators before and after the adjustment of cell resource configuration is satisfactory.

The cell data channel configuration system of the present invention further includes a cell type mapping rule set generating module, which is used to obtain the cell type mapping rule set by adopting a decision tree method according to the cell service data.

The technical effect of the present invention is that the data service channel can be reasonably planned, which not only guarantees the quality of the data service, improves user perception, but also improves the utilization efficiency of the data channel as much as possible, and improves the quality of the voice service.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention may be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

The technical solutions of the present invention will be described in further detail below with reference to the accompanying drawings and embodiments.

Description of drawings

The accompanying drawings are used to provide a further understanding of the present invention, and constitute a part of the description, and are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the attached picture:

FIG. 1 is a flow chart of a cell data channel configuration method in Embodiment 1 of the present invention;

Fig. 2 is the flowchart of the further refinement of the step of clustering to obtain the cell type in Embodiment 1 of the present invention;

FIG. 3 is a flow chart of further refinement of the steps of establishing the relationship between parameters and configurations in Embodiment 1 of the present invention;

Fig. 4 is the flow chart that the step of TRF239 prediction regression in the embodiment of the present invention is further refined;

Figure 5 is a schematic diagram of the comparison before and after the capacity reduction of TBF_38C after reconfiguring the cells that need to be reduced using the first embodiment of the present invention;

Fig. 6 is a schematic diagram of the comparison before and after the capacity reduction of TBF_16C after reconfiguring the cell that needs to be reduced using the first embodiment of the present invention;

Fig. 7 is a schematic diagram of the comparison before and after LLC_3 capacity reduction after reconfiguring the cells that need to be reduced using the first embodiment of the present invention;

Fig. 8 is a schematic diagram of the comparison before and after the single-slot load reduction after reconfiguring the cell that needs to be reduced using the first embodiment of the present invention;

Figure 9 is a schematic diagram of the comparison before and after the expansion of the TBF_38C after reconfiguring the cell that needs to be expanded using the first embodiment of the present invention;

Fig. 10 is a schematic diagram of the comparison before and after the expansion of TBF_16 after reconfiguring the cells that need to be expanded using the first embodiment of the present invention;

Fig. 11 is a schematic diagram of comparison before and after LLC_3 expansion after reconfiguring the cells that need to be expanded using Embodiment 1 of the present invention;

Fig. 12 is a schematic diagram of the comparison before and after single-slot load expansion after reconfiguring the cells that need to be expanded using the first embodiment of the present invention;

FIG. 13 is a schematic diagram of a cell data channel configuration system in Embodiment 2 of the present invention.

Detailed ways

The preferred embodiments of the present invention will be described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described here are only used to illustrate and explain the present invention, and are not intended to limit the present invention.

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

As shown in Figure 1, a kind of cell data channel configuration method comprises the steps: collecting cell service data; according to the collected cell service data, the cells with the same service characteristics are classified into one category; using regression method for each type of cell , establishing an association relationship between channel and parameter configurations and network indicators; according to the association relationship between channel and parameter configurations and network indicators for each type of cell, resource allocation is performed on corresponding cells.

Through the "data snapshot" system of Nokia Siemens, which adopts DPI (Deep Packet Inspection, deep packet identification) technology, the relevant information of the service characteristics of the cell is obtained. In order to achieve more accurate wireless resource allocation, 7*24 hours data will be collected (extract six hours of busy hours in the morning and afternoon). The service characteristics of the cell will include the data flow of various services in the cell, the frequency of the cell’s business (the number of times the service is initiated), and the duration of the cell’s business. It is required that the traffic of identifiable business types should cover at least 80% of the total traffic, at least including browsing, instant messaging, stocks, games, downloads, MMS (Multimedia Messaging Service, multimedia information service), mail and other common services, and the remaining small traffic services will no longer be subdivided.

Due to the large variety of businesses, it brings very multi-dimensional data, and it is difficult to find the required information using traditional methods. As shown in Fig. 2, the method of data mining is adopted in the method of the present invention, and the sub-districts with the same service characteristics (including information such as service distribution and service frequency) are classified into one category through a clustering algorithm. Each cell should belong to and only belong to a certain cell type.

The input data is selected and processed through data preprocessing steps. For example: missing data (ie, null values), abnormal data (not in line with normal logic or far away from most data points) that appear in the data collection process, etc., need to be excluded.

Secondly, in the live network data, the traffic of some services is very small, even in many cells, the traffic of this service is 0, so such services should be excluded from our clustering algorithm. Because there is no difference in itself, not excluding such input data will only increase the complexity of clustering and affect the results.

In addition, some data services have similar characteristics, that is, they have similar requirements on wireless channel resources, so these services can be combined into one type of service for processing.

The present invention has used two kinds of clustering algorithms of K-means and two step. Use data mining tools (such as: PASW modeler), import the preprocessed data, first use Two step to get a more reasonable number of target categories, and then set the number of target categories in the K-means algorithm, and get after each record A label of the category it belongs to, that is, the number of the desired cell type.

Verify the accuracy of the obtained cell type number. The first verification method is to verify from the evaluation criteria of the data mining method, and to measure the pros and cons of the classification through the evaluation of the cohesive and segregated contour measurements. When the index is greater than 0.25, the clustering result is good, and if it is greater than 0.5, it is good. The second method of verification is whether the obtained cell types have a high consistency for a characteristic geographical area. The third verification method is to start with the network indicators and observe whether the indicators of each cell type meet the expected service characteristics. If the result test is unqualified, re-select the appropriate variable, or adjust the parameters of the clustering algorithm, etc., until the obtained result can be accepted mathematically and operationally.

For each type of cell, the relationship between the resource configuration suitable for this type of cell and the main quality indicators is sought. As shown in Figure 3, the process of establishing the relationship between configuration and parameters and network indicators in the regression method is divided into three steps: calculating TRF239; calculating CDEF; and calculating CDED. By calculating TRF239, establish a prediction model for network index TRF239; by calculating TRF239, establish a prediction model for configuration parameter CDEF; by calculating TRF239, calculate the corresponding CDED parameter configuration according to the congestion rate.

TRF_239 counts the number of time slots actually used for downlink data services, reflecting the actual use of radio resources for data services. Within a specific index selection range, it can be regarded as the required number of data wireless resources. Studying the corresponding relationship between TRF_239 and the composition of various services and other indicators in different clustered cells can be used as an evaluation criterion for the later service growth or wireless resource requirements of new sites, and facilitates the calculation and planning of resource requirements.

As shown in Figure 4, the establishment of a prediction model for the network indicator TRF239 mainly includes the following steps:

Selected predictive independent variable; perform data conversion on the independent variable; use the converted data of the independent variable as input, use the algorithm of multiple linear regression to calculate the result; analyze the result, if the evaluation index cannot be reached, re-select Independent variables, or perform other data conversion on the independent variables, or choose other fitting algorithms, until the evaluation index is reached.

About selecting key indicator parameters. Combined with business knowledge and correlation analysis, the final selected predictive variables are: TBF_38C (Average DL TBF per timeslot, the average downlink TBF per time slot, TBF is the abbreviation of Temporary Block Flow), RLC_DL (Radio LinKControl_DownLinK, data Downlink traffic), the frequency of various services in the community, and the proportion of traffic.

Data transformation is performed on the independent variable so that there is a linear relationship between the transformed data and the target variable. Available transformations include logarithmic, quadratic, cubic, inverse, power, S, growth, exponential, and more. In one embodiment of the method of the invention, power transformation is used for TBF_38C.

On the choice of multiple linear regression algorithm. There are four types of algorithm parameters, namely: Enter, Step, Forward and Backward. In one embodiment of the method of the present invention, Step is selected for fitting, and for each type of cell, the input variables selected by the algorithm are different.

On the analysis of fitting results. When the fitting equation is obtained, some main indicators for evaluating the fitting accuracy can also be obtained at the same time, such as R2, etc. If the goodness of fit is 0.80, the accuracy requirement is met. If the evaluation index cannot be reached, then the independent variable needs to be reselected, or the independent variable should be processed, or another fitting algorithm should be selected.

CDEF refers to the channel used for data services by default, that is, the data dynamic channel. When the voice service is busy and the voice channel occupies a certain proportion, it can be converted to a voice channel for use. In the method of the present invention, CDEF includes CDED static time slot, so CDEF>=CDED. A reasonable CDEF setting value will enable the cell to reduce the number of lifting domains and reduce the system load of the system.

After obtaining TRF_239, use it as an independent variable, and at the same time take the network configuration parameters of the lifting domain, cell voice single time slot load and data single time slot load as independent variables, and use the linear regression method to establish the correlation between it and CDEF , to get a reasonable CDEF configuration value.

CDEF is the boundary of the default CS (Circuit Service, voice domain) and PS (Packet Service data domain) domains, and its settings have the following characteristics:

It will not be frequently upgraded due to the impact of PS traffic in the PS domain;

It will not be frequently downgraded due to the impact of CS traffic in the CS domain;

Because the setting of the CS traffic load algorithm is changed, the change of CDEF will not affect the half-rate ratio of CS;

Because CDEF will not affect TRF_239 and data traffic, it will not cause considerable changes in the efficiency of PDCH (Packet Data Channel, packet data channel).

CDED refers to a channel dedicated to data services, that is, a static data channel, which cannot be occupied by voice services under any circumstances and is only allocated to data services. If CDED is set too large, it will not be able to occupy more resources in the data domain when the voice traffic is high, which may cause congestion. And if CDED is set too small, then in the case of high voice traffic, the voice traffic will squeeze the data dynamic channel until the static channel, and too small static signal state setting value will cause data traffic congestion.

The setting of CDED is to ensure the use of data services at a minimum, so the setting value of CDED should study the minimum channel configuration value without causing hard data congestion in the cell. In the method of the present invention, the hard congestion index TBF_16 is selected from the range of 0.5%-1% as the threshold for hard congestion.

Combined with the voice service of the cell, calculate the number of channels required for voice (according to the currently widely used ERL-B algorithm) and the number of channels required for data services (according to the results of data channel resource requirements studied earlier), to give the growth rate of the cell. Capacity or capacity reduction suggestions, and configure new CDED/CDEF (data static time slot/data dynamic time slot) parameters according to the new carrier frequency input plan.

Record and observe the changes in KPI (Key Performance Indicator, key performance indicator method) of the week before and after the adjustment of the resource configuration of the community, including data and voice. If the index changes are unsatisfactory or cannot be explained, it is necessary to consider adjusting the parameter selection and re-mining the relationship between parameters and network indicators. In an actual project, the cell data channel configuration method of the present invention is used to adjust the configuration of the corresponding cell according to the obtained allocation information of each cell. For the cells that need to reduce the allocation, the before and after comparison of the indicators is wrong! Reference source not found. As shown in Figure 5 to Figure 8 (for each pair of comparison squares that are close to each other, the square on the left shows before adjustment, and the square on the right shows after adjustment), it can be seen that although the key indicators have deteriorated slightly, they all conform to the network quality Requirements for index assessment. In this way, the purpose of maintaining service quality, saving channels and improving network resource utilization efficiency is achieved. For a community that needs to be expanded, its key indicators are shown in Figures 9 to 12 (for each pair of comparison squares that are close to each other, the square on the left indicates before adjustment, and the square on the right indicates after adjustment), all of which have been improved to reach It meets the requirements of index assessment, improves the quality of data services, and enhances user perception.

Generate cell channel planning principles and integrate the principles into software for subsequent large-scale deployment. The interface design of the tool software is required to be concise and clear, and the operation is simple and easy to use. It needs to have an interface with the database, and the required data tables can be obtained from the existing database. The function needs to meet data import, browsing, analysis, configuration automatic generation, result export, etc.

The cell type mapping rules are obtained by using the decision tree method. After the cell type is obtained through the clustering method according to the cell service information data, in order to be popularized and used, it is necessary to be able to determine the service characteristics of each cell and the rules of the cell type it belongs to. In this way, as long as there is new data input, the cell type to which it belongs can be obtained immediately.

Using the decision tree method, by training the service characteristic data of the cell, find out the corresponding relationship between it and the cell type. In one embodiment of the method of the present invention, the decision tree algorithm of C5.0 is adopted to realize effective division among all cells and available cell types.

According to the business data of the community, the decision tree method in the data mining tool is used to obtain the mapping rule set of the community type, and the corresponding relationship between the service characteristic data of the community and the type of the community is recorded, and the rule set is all expressed in the form of If...then...else . Therefore, in the tool, the mapping from the input cell feature information to the cell type is realized first through the solidified rule set. Then, through the previously obtained regression model suitable for each type of cell, the reasonable channel configuration parameters suitable for this type of cell and for different service flows of each cell are obtained.

The present invention also discloses a community data channel configuration system, as shown in Figure 13, the system includes: a data collection module, a community classification module and a configuration module; a data collection module for collecting community business data; a community classification module for Based on the cell service data collected by the data collection module, the cells with the same service characteristics are classified into one category; the configuration module is used to establish corresponding channel and parameter configurations and network for each type of cell divided by the cell classification module Relationships between indicators.

A cell data channel configuration system, further comprising an evaluation module, which is used to evaluate the resource configuration before and after adjustment of the resource configuration of the corresponding cell according to the association relationship between the channel and parameter configuration for each type of cell and the network index If the change of key performance indicators is unsatisfactory, call the configuration module to re-establish the correlation between channel and parameter configuration and network indicators, until the evaluation result of the change of key performance indicators before and after the adjustment of cell resource configuration is satisfactory.

A cell data channel configuration system further includes a cell type mapping rule set generating module, which is used to obtain the cell type mapping rule set by adopting a decision tree method according to cell service data.

To sum up, in the present invention, by adopting the method and system for configuring a cell data channel, the data service channel can be reasonably planned, which not only guarantees the quality of the data service, improves user perception, but also maximizes the user experience as much as possible. Improve the utilization efficiency of data channels and improve the quality of voice services.

Claims (14)

1. a cell data channel arrangement method, is characterized in that, comprises step:

Gather the Zone data;

According to the Zone data that gather, the residential quarter that traffic performance is identical is classified as a class;

Adopt homing method for the residential quarter of each type, set up the incidence relation between channel and parameter configuration and network index;

According to for the channel of the residential quarter of each type and the incidence relation between parameter configuration and network index, respective cell is carried out resource distribution.

2. the method for claim 1, is characterized in that, and is further comprising the steps of:

Evaluate and test described respective cell resource distribution and adjust the variation of front and back KPI Key Performance Indicator, if evaluation result is undesirable, rebulid the incidence relation between channel and parameter configuration and network index, to desirable to described evaluation result.

3. method as claimed in claim 2, is characterized in that, obtains the KPI of local resource configuration, judges whether described KPI reaches preset value, if reach, described evaluation result is desirable, otherwise described evaluation result is undesirable.

4. want 1,2 or 3 described methods as right, it is characterized in that, the Zone data gather by the data snapshot system.

5. want 1,2 or 3 described methods as right, it is characterized in that, described traffic performance comprises data traffic, the Zone frequency and the Zone duration of business.

6. as claim 1,2 or 3 described methods, it is characterized in that, adopt the method for data mining, by clustering algorithm, the residential quarter that traffic performance is identical is classified as a class.

7. method as claimed in claim 6, further comprise,

The Zone data that gather are carried out preliminary treatment, the Zone data after being processed;

As input, first utilize clustering algorithm Two step to obtain target classification number with the Zone data after processing, take target classification number as input, utilize clustering algorithm K-means to obtain the cell type numbering;

The cell type numbering is verified, if the result is undesirable, adjusted the parameter of Two step and/or K-means, until result is desirable.

8. as claim 1,2 or 3 described methods, other are characterised in that, the step of setting up the incidence relation between channel and parameter configuration and network index further comprises,

The forecast model of foundation to network index TRF239;

The forecast model of foundation to configuration parameter CDEF;

Calculate corresponding CDED parameter configuration according to congestion ratio.

9. method as claimed in claim 8, is characterized in that, the step of setting up the forecast model of network index TRF239 further comprises,

Selected prediction independent variable;

Independent variable is carried out data transaction;

Data after changing take independent variable are used the algorithm of multiple linear regression as input, calculate result;

Result is analyzed, if do not reach evaluation index, reselected independent variable, perhaps independent variable is carried out the data transaction of alternate manner, perhaps select other fitting algorithm, to reaching evaluation index.

10. method as claimed in claim 8, is characterized in that, the step of setting up the forecast model of configuration parameter CDEF further comprises,

TRF_239, lifting territory network configuration parameters, residential quarter speech list time-slot load and data sheet time-slot load as independent variable, are used linear regression algorithm, calculate CDEF.

11. as claim 1,2 or 3 described methods, it is characterized in that, also comprise:

According to the Zone data, adopt traditional decision-tree to obtain cell type mapping ruler collection.

12. a cell data channel configuration system is characterized in that, comprising: data acquisition module, cell classification module and configuration module;

Data acquisition module is used for gathering the Zone data;

The cell classification module, for the Zone data that gather according to data acquisition module, the residential quarter that traffic performance is identical is classified as a class;

Configuration module for the residential quarter of each type of dividing for the cell classification module, is set up corresponding channel and the incidence relation between parameter configuration and network index.

13. device as claimed in claim 12, it is characterized in that, also comprise the evaluation and test module, be used for basis for the channel of the residential quarter of each type and the incidence relation between parameter configuration and network index, after respective cell is carried out resource distribution, the variation of KPI Key Performance Indicator before and after the configuration of evaluation and test local resource is adjusted, if undesirable, call configuration module and rebulid incidence relation between channel and parameter configuration and network index, desirable to the result to the variation evaluation and test of KPI Key Performance Indicator before and after configuration is adjusted to local resource.

14. system as described in claim 12 or 13 is characterized in that, also comprises cell type mapping ruler collection generation module, is used for according to the Zone data, adopts traditional decision-tree to obtain cell type mapping ruler collection.

Images