Method and system for automatically identifying key event from sales session

Abstract

The invention relates to a method and a system for automatically identifying a key event from a sales session, belongs to the technical field of natural language processing, and solves one or more defects of an existing key event identification method. The method comprises the following steps: acquiring an event rule of each to-be-identified key event; segmenting the sales session to obtain a plurality of text-format session sentences identifying the session role; taking the session sentences containing the keywords as sentences to be recognized, and preliminarily matching each event rule with each sentence to be recognized to obtain a plurality of sentences to be recognized preliminarily matched with each event rule; for each preliminarily matched event rule, judging whether the event rule contains other options or not, and if not, identifying the to-be-identified statement preliminarily matched with the event rule as a key event; and if yes, matching the event rule with the to-be-identified statement preliminarily matched with the event rule again, and identifying the to-be-identified statement passing the matching again as a key event.

Description

technical field [0001] The present invention relates to the technical field of natural language processing, in particular to a method and system for automatically identifying key events from sales conversations. Background technique [0002] A sales conversation includes a series of key events, marking the product attributes introduced by the sales, the customer's attitude towards the product, the salesman's words in response to customer objections, whether to mention competing companies, whether to successfully agree on the next step, etc. Identifying key events from a large number of conversations is of great help in analyzing conversation quality, evaluating customer intentions, discovering business opportunities and risks, reviewing sales problems, and refining excellent cases. [0003] The efficiency of manually identifying key events is low, and it cannot cope with the analysis and processing of massive sessions. The existing key event identification technologies are ...

AI Technical Summary

Problems solved by technology

[0006] Second, keyword matching has relatively large limitations, resulting in low accuracy of key event identification

[0008] Third, the classification model is less efficient in actual use

If the natural sentences in the dialogue are used as the granularity and input into the model sentence by sentence for judgment, the calculation efficiency is relatively low, and if different context pairs are considered, the calculation scale will increase at the quadratic level, which cannot be applied in the actual production environment

Even if the model is only iteratively updated for a few key events, when it is desired to re-optimize and analyze a large amount of historical data, all events will be affected together, and the full amount of data still needs to be run, which is relatively inefficient

[0010] In summary, the traditional key event identification method has defects such as low accuracy and recall rate, low operating efficiency and poor practical application effect.

Images