Large data analytical approach and terminal thereof based on investigating in error
Technical field
The present invention relates to large data analysis and applied technical field, relate more specifically to a kind of large data analytical approach and terminal thereof based on investigating in error.
Background technology
At present, the analytic product of large data mostly adopts raw data Direct Sampling, afterwards raw data is sent to data analytics server and carries out analyzing and processing.In raw data, have a large amount of repetitions, tediously long data, and the data of these repetitions are often contributed not quite to analysis result, or even do not had resultful operation.Therefore, directly raw data is sent to data analytics server, because data are huge, huge computational burden will be brought to data analytics server, and along with the increase day by day of data, often cause hardware device to catch up with, when serious, can directly have influence on precision of analysis.In addition, directly raw data is sent to data analytics server, raw data is not carried out to any error investigation, understand equally the accuracy of impact analysis result.
Therefore, be badly in need of a kind of improved large data analytical approach and terminal thereof based on investigating in error and overcome above-mentioned defect.
Summary of the invention
The object of this invention is to provide a kind of large data analytical approach based on investigating in error, to reduce the analytic operation amount of data analytics server to large data, improve precision of analysis.
Another object of the present invention is to provide a kind of large data analysing terminal based on investigating in error, to reduce the analytic operation amount of data analytics server to large data, improves precision of analysis.
For achieving the above object, the invention provides a kind of large data analytical approach based on investigating in error, comprising:
Reception is from the enabled instruction of data analytics server;
According to described enabled instruction, the input data from data platform are carried out investigating in error;
According to investigation result, optionally the sampling time label of described input data or described input data is sent to described data analytics server.
Compared with prior art, method of the present invention carries out investigating in error to the input data from data platform according to the enabled instruction of data analytics server, and optionally input data are sent to data analytics server according to investigation result, thereby reduced the analytic operation amount of data analytics server to large data, improved precision of analysis.
Preferably, the enabled instruction receiving from data analytics server also comprises before:
Reception is from the input data of described data platform.
Preferably, the input data that receive from described data also comprise afterwards:
Reception is from error judgment value and the data model of described data analytics server.
Particularly, according to investigation result, optionally the sampling time label of described input data or described input data being sent to described data analytics server specifically comprises:
Judge whether described investigation result is greater than described error judgment value;
When described investigation result is greater than described error judgment value, described input data are sent to described data analytics server;
When described investigation result is less than or equal to described error judgment value, the sampling time label of described input data is sent to described data analytics server.
Preferably, according to investigation result, optionally described input data being sent to described data analytics server also comprises afterwards:
Described data analytics server is according to error judgment value and data model described in described input Data Update, and described error judgment value after renewal and error model are outwards transmitted.
Correspondingly, the present invention also provides a kind of large data analysing terminal based on investigating in error, comprising:
Instruction module, for receiving the enabled instruction from data analytics server;
Investigation module, for carrying out investigating in error to the input data from data platform according to described enabled instruction;
Output module, for being optionally sent to described data analytics server by the sampling time label of described input data or described input data according to investigation result.
Preferably, this terminal also comprises:
Load module, for receiving error judgment value and the data model from the input data of described data platform, described data analytics server.
Preferably, this terminal also comprises:
Judge module, for judging whether investigation result is greater than described error judgment value.
Preferably, this terminal also comprises:
Sampling module, for extract the sampling time label of the input data of the described load module of input according to the investigation result of described investigation module, and is sent to described output module by described sampling time label.
By following description also by reference to the accompanying drawings, it is more clear that the present invention will become, and these accompanying drawings are used for explaining embodiments of the invention.
Accompanying drawing explanation
Fig. 1 is the process flow diagram that the present invention is based on large data analytical approach one embodiment of investigation in error.
Structured flowchart when Fig. 2 applies for the large data analysing terminal based on investigating in error.
Fig. 3 is the theory diagram of Fig. 2.
Embodiment
With reference now to accompanying drawing, describe embodiments of the invention, in accompanying drawing, similarly element numbers represents similar element
Please refer to Fig. 1, the large data analytical approach that the present invention is based on investigation in error comprises the following steps:
S101, receives the input data P from data platform
i(data
1, data
2data
n);
S102, receives error judgment value and data model from data analytics server; Particularly, data analytics server completes error judgment value e and data model [e, P according to the data that receive before
s(data
1, data
2data
n)] structure;
S103, receives the enabled instruction Command[start from data analytics server]; It should be noted that data analytics server can according to before the Data Update error judgment value and the data model that receive, when error judgment value and data model upgrade to some extent, data analytics server will outwards send update instruction Command[update];
S104, carries out investigating in error computing Delta[(P according to enabled instruction to input data
s(data
1, data
2data
n), P
i(data
1, data
2data
n)];
S105, whether judgement investigation result is greater than error judgment value, if so, carries out S106, otherwise, carry out S107;
S106, is sent to data analytics server by input data;
S107, is sent to data analytics server by the sampling time label of input data, or does not export any data.
Correspondingly, please refer to Fig. 2 and Fig. 3, the present invention also provides a kind of large data analysing terminal based on investigating in error, comprising:
Load module 10, for receiving the input data from data platform, error judgment value and the data model of data analytics server;
Instruction module 11, for receiving the enabled instruction from data analytics server; It should be noted that data analytics server can according to before the Data Update error judgment value and the data model that receive, when error judgment value and data model upgrade to some extent, instruction module receives the update instruction of data analytics server;
Investigation module 12, for carrying out investigating in error to input data according to enabled instruction;
Judge module 13, for judging whether investigation result is greater than error judgment value;
Sampling module 14, for extracting the sampling time label of input data, and transmits sampling time label;
Output module 15, for being optionally sent to data analytics server by the sampling time label of input data or input data according to investigation result.
Particularly, when judge module is judged investigation result while being greater than error judgment value, indication output is touched input data is sent to data analytics server, otherwise, indicate sampling module to extract the sampling time label of input data, and sampling time label is sent to output module, and by output module, sampling time label is sent to data analytics server.
As can be seen from the above description, the present invention is based on large data analytical approach and the terminal thereof of investigation in error, there is following beneficial effect:
(1) particularly for the higher plateform system of repeated data, because the method and terminal carry out investigating in error to the input data from data platform according to the enabled instruction of data analytics server, and optionally input data are sent to data analytics server according to investigation result, thereby reduced the analytic operation amount of data analytics server to large data, improved precision of analysis;
(2) data analytics server is according to input Data Update error judgment value and data model, and error judgment value after renewal and error model are outwards transmitted, be dynamically alignment error value e and data model Ps of data analytics server, so can dynamically adjust operand, greatly strengthen dirigibility and the conformability of data analytics server (system) self.
(3) strengthened the trace ability of historical data, when applications client whenever necessary, can provide strong data history to case study, investigation operating personnel.
(4) for the data of each different platform, the present invention can adopt for the personality data of these platforms different error computational algorithms (Delta), has effectively improved algorithm dirigibility and diversity for whole huge data analysis.
Invention has been described for above combination most preferred embodiment, but the present invention is not limited to the embodiment of above announcement, and should contain the various modifications of carrying out according to essence of the present invention, equivalent combinations.