FR3126797A1 - Process for the exploitation of collected data, electronic device and corresponding computer program and support products - Google Patents

Abstract

Data exploitation method, device, computer program product and corresponding medium A method comprising: obtaining contextual data relating to electronic messages sent or received by a plurality of communication terminals, contextual data comprising at least: - a caption of an electronic message; - at least one identifier of a sender and a receiver of said message; - a timestamp of said message; obtaining a first contextual datum and/or an identifier of a first sender of a first contextual datum, according to a similarity between labels of said contextual data and a similarity between values of a descriptive information of said contextual data and a first value of said descriptive information. The invention also relates to the corresponding device, computer program product and medium. Figure for the abstract: Fig. 3

Description

Process for the exploitation of collected data, electronic device and computer program product and corresponding medium

1. Technical area

This application relates to the field of exploitation, at least partially automatic, of data concerning at least one group of individuals. This may include a large group of individuals.

It relates in particular to a process for exploiting such data as well as electronic devices, computer program products and corresponding media.

2. State of the art

Knowing the members of a group is an issue for the people in charge of managing this group, especially when the size of the group grows.

For example, identifying actors with influence within the group (also called influencers in this application) can help spread information within the group based on these actors. For example, in the context of the deployment of a new technical solution in the group, convincing these actors of the benefits provided by this solution can help the acceptance of the solution within the group. In addition, training these actors as a priority in such a solution can contribute to the understanding and handling by the whole group of this solution. Therefore, the identification of influencers within a company can be a management tool for human resources departments or decision-makers.

Studies have been proposed to identify such actors within social networks, by studying the spread of information in these social networks. These studies use notions that are generally very specific and specific to membership and to the codes of the social network, which cannot easily be translated into the field of business (such as notions such as tweets, hashtags, or followers, better known respectively, according to the English terminology, under the name of tweets, hashtag, followers).

Also, these studies are difficult to apply within the framework of a real structure of individuals, such as a company, a trade union or non-governmental organization, due to the diversity of the actors (who are not all members of a social network) and the private nature of the information exchanged within the structure. Thus, there is a need for a solution making it possible to analyze information exchanged within a group of individuals of a real structure, such as the members of a company.

The purpose of this application is to propose improvements to at least some of the drawbacks of the state of the art.

3. Disclosure of Invention

The present application aims to improve the situation, using a method comprising:

obtaining a plurality of contextual data relating to electronic messages transmitted or received by a plurality of terminals of a communication network, a contextual data of said plurality of contextual data comprising at least:

- a description of an electronic message;

- at least one identifier of at least one sender and/or receiver of said message;

- a timestamp of said message;

obtaining at least one first contextual data item from said plurality of contextual data items and/or at least one identifier of a first sender of at least one first contextual data item from said plurality of contextual data items, as a function of a similarity between labels of said contextual data and a similarity between at least one piece of information descriptive of said contextual data with at least one first value.

According to at least certain embodiments, said obtaining of said identifier of said first sender and/or of said first contextual data comprises a grouping of said contextual data according to a similarity between their labels.

According to at least certain embodiments, said first value is associated with a standard profile of senders or receivers of said electronic messages.

According to at least certain embodiments, said obtaining of said identifier of said first transmitter and/or of said first contextual data comprises chaining at least a portion of at least two of said grouped contextual data, said chaining taking into account the transmitters and/or or receivers and/or timestamps of said aggregated contextual data.

According to at least certain embodiments, said chaining takes account of the senders and receivers of said grouped contextual data.

According to at least certain embodiments, said chaining takes account of portions of said wordings indicative of a propagation, between at least two of said transmitters and/or receivers of said grouped contextual data, of said electronic messages to which said grouped contextual data relate.

According to at least certain embodiments, said obtaining of said identifier of said first sender and/or of said first contextual data comprises a normalization of the labels of said contextual data producing standardized labels, said normalization comprising a deletion of said portions indicative of said labels of said contextual data, said grouping taking into account said standardized wordings.

According to at least certain embodiments, said descriptive information relates to said grouped contextual data.

According to at least certain embodiments, the timestamp of said second contextual data is earlier than the timestamps of said contextual data grouped with said second contextual data.

According to at least certain embodiments, said descriptive information belongs to a group comprising:

• a number of senders and/or receivers of said second contextual datum;
• a number of senders and/or receivers other than the senders and/or receivers of said second contextual datum in said at least one of said contextual data grouped with said second contextual datum;
• a total number of contextual data grouped with said second contextual data;
• a percentage of contextual data with the same label and the same timestamp as said second contextual data, in said at least one of said contextual data grouped with said second contextual data, relative to said total number;
• a percentage of contextual data relating to a response message to another message, in said at least one of said contextual data grouped with said second contextual data, relative to said total number;
• a percentage of contextual data relating to a transfer message of another message, in said at least one of said contextual data grouped with said second contextual data, relative to said total number.

According to at least certain embodiments where said messages are emails, said descriptive information belongs to a group comprising:

• a number of senders and/or receivers of a creation email associated with at least one of said grouped contextual data.
• a number of senders and/or receivers other than the senders and/or receivers of said creative email in said at least one of said aggregated contextual data associated with said creative email;
• a total number of aggregated context data associated with said creative email;
• a percentage of creative email context data in said aggregated context data associated with said creative email, relative to said total number;
• a percentage of response email context data in said aggregated context data associated with said creation email, relative to said total number;
• a percentage of forwarding email context data in said aggregated context data associated with said creation email, out of said total number.

According to at least certain embodiments, said descriptive information is representative of a propagation of an electronic message to which at least one second of said grouped contextual data relates, between said senders and/or receivers of said grouped contextual data with said second contextual data .

According to at least certain embodiments, the method comprises an anonymization of said contextual data.

Said method can be implemented at least partially in an electronic device, for example a communication device belonging to said communication network.

In at least one embodiment, said sender and/or receiver is at least one user of at least one terminal of said communication network.

The characteristics, presented separately in the present application in connection with certain embodiments of the method of the present application can be combined with each other according to other embodiments of the method.

According to another aspect, the present application also relates to an electronic device suitable for implementing the method of the present application in any one of its embodiments.

For example, the present application thus relates to an electronic device comprising at least one processor configured to:

obtaining a plurality of contextual data relating to electronic messages transmitted or received by a plurality of terminals of a communication network, a contextual data of said plurality of contextual data comprising at least:

- a description of an electronic message;

- at least one identifier of at least one sender and at least one receiver of said message;

- a timestamp of said message;

obtaining at least one first contextual data item from said plurality of contextual data items and/or at least one identifier of a first sender of at least one first contextual data item from said plurality of contextual data items, as a function of a similarity between labels of said contextual data and a similarity between values of at least one descriptive item of said contextual data and at least one first value of said descriptive item.

The present application also relates to a computer program comprising instructions for the implementation of the various embodiments of the above method, when the program is executed by a processor and a recording medium readable by an electronic device and on which the computer program is recorded.

For example, the present application thus relates to a computer program comprising instructions for the implementation, when the program is executed by a processor of an electronic device, of a method comprising:

obtaining a plurality of contextual data relating to electronic messages transmitted or received by a plurality of terminals of a communication network, a contextual data of said plurality of contextual data comprising at least:

- a description of an electronic message;

- at least one identifier of at least one sender and at least one receiver of said message;

- a timestamp of said message;

obtaining at least one first contextual data item from said plurality of contextual data items and/or at least one identifier of a first sender of at least one first contextual data item from said plurality of contextual data items, as a function of a similarity between labels of said contextual data and a similarity between values of at least one descriptive item of said contextual data and at least one first value of said descriptive item.

Furthermore, the present application relates for example to a recording medium readable by a processor of an electronic device and on which is recorded a computer program comprising instructions for the implementation, when the program is executed by the processor , of a method comprising:

obtaining a plurality of contextual data relating to electronic messages transmitted or received by a plurality of terminals of a communication network, a contextual data of said plurality of contextual data comprising at least:

- a description of an electronic message;

- at least one identifier of at least one sender and at least one receiver of said message;

- a timestamp of said message;

obtaining at least one first contextual data item from said plurality of contextual data items and/or at least one identifier of a first sender of at least one first contextual data item from said plurality of contextual data items, as a function of a similarity between labels of said contextual data and a similarity between values of at least one descriptive item of said contextual data and at least one first value of said descriptive item.

The information carriers mentioned above can be any entity or device capable of storing the program. For example, a medium may comprise a storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or even a magnetic recording means.

Such a storage means can for example be a hard disk, a flash memory, etc.

On the other hand, an information medium can be a transmissible medium such as an electrical or optical signal, which can be conveyed via an electrical or optical cable, by radio or by other means. A program according to the invention can in particular be downloaded from an Internet-type network.

Alternatively, an information medium may be an integrated circuit in which a program is incorporated, the circuit being adapted to execute or to be used in the execution of any one of the embodiments of the method which is the subject of the present application for patent.

4. Brief description of drawings

Other characteristics and advantages of the invention will appear more clearly on reading the following description of particular embodiments, given by way of simple illustrative and non-limiting examples, and the appended drawings, among which:

There presents a simplified view of a system, cited by way of example, in which at least certain embodiments of the method of the present application can be implemented,

There presents a simplified view of a device suitable for implementing at least certain embodiments of the method of the present application,

There outlines the method of the present application, in some of its embodiments

There presents an example representation of a creative email;

There shows an example representation of a reply email;

There presents an example representation of a forwarding email;

There presents an example representation of a reply-and-forward email;

There presents an example of a directed graph representing a communication chain and the associated transfer graph, in certain embodiments of the method of the present application;

There presents an example of weighting the transfer graph of the .

5. Description of embodiments

The invention lies in the field of information exchanges within a group of individuals. The group of individuals may in particular comprise a large number of individuals. They may be, for example, members of a public or private company, an organization, an association, a sports federation, and/or a community, having for example a large number of employees, customers and/or suppliers (or more generally interlocutors).

The invention proposes to analyze data relating to electronic messages (or data associated with these messages) exchanged within the group of individuals in order to deduce therefrom properties associated with at least one of the individuals of the group.

In certain embodiments, the analysis may not take into account the messages sent from this group of individuals to the outside of the group and the messages sent from outside the group to this group of individuals. In certain embodiments, on the contrary, the analysis can take into account messages sent from this group of individuals to outside the group and/or messages sent from outside the group to this group of individuals. For example, in certain embodiments, it is possible to consider the messages sent by an individual of the group and including in their addressee(s) at least one addressee outside the group, without taking this external addressee into consideration, or taking into account this external addressee but ignoring the messages transmitted in turn by this external addressee (without therefore being interested, for example, in the propagation of the message outside the group). In certain embodiments, it is possible for example to take into account the messages transmitted by a sender outside the group but taking into account only the reactions of the individuals of the group to this message, (and not the possible reactions of other recipients outside the group to this message).

Data associated with messages used for analysis is referred to in this application as contextual data. In particular, they can help to characterize the exchanges. The contextual data of a message can include, for example, elements characterizing the sending of a message such as a date and/or time of sending of the message, addresses and/or identifiers of sender and recipient(s) of the message, and/or information representative of the semantics of the message, such as a label (from, for example, the title or the body of the message) (for example, in the case where the message is an email, a subject of this email) . Using, in some embodiments, contextual data from a message instead of the message itself can help limit the size of the information handled by the method of the present application, and therefore the memory occupation and the time of associated treatment. In addition, the use of a label instead of the entire content of a message helps to preserve a certain confidentiality of the messages considered.

Electronic message context data may be grouped into strings, also referred to as "communication strings" (and in some cases "propagation strings") in this application. The properties of these communication chains can be used for example to detect, among the individuals of the group, individuals of at least one first type, by similarity with first characteristics (constituting a standard profile) associated with this first type of people. For example, in embodiments analyzing a propagation of information, first characteristics can be associated with a type of individuals of the “influencer” type.

The electronic messages (or the associated contextual data) can for example be collected automatically (or at least partially automatically) concerning individuals of the group to constitute or enrich a knowledge base associated with the group, which is then exploited according to the present application. These may be messages such as audio, video and/or textual electronic messages such as electronic mail (email or "email" in English terminology) or instant messages (chat). These electronic messages can for example be chosen so as to be representative of the actual exchanges within the group. For example, in certain embodiments adapted to a group in which the majority of the exchanges are based on a single type of electronic message (emails for example), the analysis can only consider messages of this type (emails in example above).

For example, in a company where communication is based on the use of emails, all emails exchanged can be considered representative of the activity of all individuals, unlike exchanges via other collaborative tools, less widespread in the company (for example used only by a small number of individuals).

Therefore, their analysis can help to understand how information is exchanged in the group (for example its propagation) and to detect specific characteristics, such as propagation characteristics used to detect influencers.

In certain embodiments, all the messages collected (or the contextual data associated with these messages) can be exploited. In other embodiments, only a portion of these messages (or associated contextual data) can be used. These may in particular be messages transmitted during a time window. Such an embodiment may for example be better suited to the use of collections of large size (such as a “log” file bringing together billions of messages transmitted over a period of several years). In addition, certain properties of individuals that may have an influence on the messages exchanged are likely to change over time (for example, the functions and/or the entities to which individuals are attached within the group may change). The narrower the time window to which the collection relates, the lower the probability of variation of certain properties of individuals in this window, thus promoting stability and therefore consistency of the data and/or messages to be exploited. This time window can for example be chosen from the order of one to a few weeks or from the order of one to a few months (for example a window for sending messages of one month or 6 months).

The present invention is presented in more detail below in relation to embodiments, cited by way of non-limiting examples, which relate to a knowledge base comprising contextual data concerning individuals of a company. Individuals are, for example, employees of the company or part of a company. The contextual data relates to messages of the “email” type.

Thereafter, “email” will designate either the e-mail itself, or for simplicity the contextual data of this e-mail, the meaning of the term appearing obvious, in the context of its use, to a person skilled in the art.

We now describe, in connection with the , in more detail this application.

There depicts a telecommunications system 100 in which certain embodiments of the invention may be implemented. The system 100 comprises one or more electronic devices, at least some of which can communicate with each other via one or more communication networks 120, possibly interconnected, such as a local network or LAN (Local Area Network, 1) and/or a wide area type network , or WAN (Wide Area Network, 2). For example, the network may include a corporate or home LAN network and/or a WAN network of the internet or cellular type, GSM - Global System for Mobile Communications, UMTS - Universal Mobile Telecommunications System, Wifi - Wireless, etc.) .

As illustrated in , the system 100 can also comprise several electronic devices, such as a terminal (such as a portable computer 110, a smartphone 130, 132, 134, a tablet 136), and/or a server 140, for example an application server , a storage device 150. The system may also include management and/or network interconnection elements (not shown). These electronic devices can be associated with at least one individual 160, 162, 164, 166 (through, for example, a user account accessible by login or a configuration of a messaging tool running at least partially on this device, several of the electronic devices 110, 130 being able to be associated with the same user 160.

There illustrates a simplified structure of an electronic device 200 of the system 100, for example the device 110, 130, 132,134, 136 or 140 of the , adapted to implement the principles of this application. According to the embodiments, it may be a server, and/or a terminal.

The device 200 notably comprises at least one memory M 210. The device 200 can notably comprise a buffer memory, a volatile memory, for example of the RAM type (for "Random Access Memory" according to the English terminology), and/or a non-volatile memory. volatile (for example of the ROM type (for "Read Only Memory" according to the English terminology). The device 200 can also comprise a processing unit UT 220, equipped for example with at least one processor P 222, and controlled by a program computer PG 212 stored in memory M 210. On initialization, the code instructions of the computer program PG are for example loaded into a RAM memory before being executed by the processor P. Said at least one processor P 222 of the processing unit UT 220 can in particular implement, individually or collectively, any of the embodiments of the method of the present application (described in particular in relation to the ), according to the instructions of the computer program PG.

The device can also comprise, or be coupled to, at least one I/O input/output module 230, such as a communication module, allowing for example the device 200 to communicate with other devices of the system 100, via wired or wireless communication interfaces, and/or such as an interfacing module with a user of the device (also called more simply in this application “user interface”). an interface integrated into the device 200, or part of a third-party device coupled to this device by wired or wireless communication means. For example, it may be a secondary screen of the device or a set of loudspeakers connected by wireless technology to the device.

A user interface can in particular be a so-called “output” user interface, suitable for rendering (or controlling a rendering) of an output element of a computer application used by the device 200, for example an application running at least partially on the device 200 or an "online" application running at least partially remotely, for example on the server 140 of the system 100. English terminology) on at least one user interface, in any form, for example comprising textual, audio and/or video components, or a combination of such components.

Examples of output user interface of the device include one or more screens, in particular at least one graphic screen (touchscreen for example), one or more loudspeakers, a connected headset.

Furthermore, a user interface can be a so-called “input” user interface, suitable for acquiring information from a user of the device 200. It can be in particular information intended for an application computer accessible via the device 200, for example an application running at least partially on the device 200 or an "online" application running at least partially remotely, for example on the server 140 of the system 100. Examples of The input user interface of the device 200 includes a sensor, an audio and/or video acquisition means (microphone, camera (webcam) for example), a keyboard, a mouse.

In certain embodiments, said at least one microprocessor of device 200 can for example be adapted to:

obtaining a plurality of contextual data relating to electronic messages transmitted or received by a plurality of terminals of a communication network, a contextual data of said plurality of contextual data comprising at least:

- a description of an electronic message;

- at least one identifier of at least one sender and at least one receiver of said message;

- a timestamp of said message;

obtaining at least one first contextual data item from said plurality of contextual data items and/or at least one identifier of a first sender of at least one first contextual data item from said plurality of contextual data items, as a function of a similarity between labels of said contextual data and a similarity between values of at least one descriptive item of said contextual data and at least one first value of said descriptive item.

Some of the above input-output modules are optional and may therefore be absent from device 200 in some embodiments. In particular, if the present application is sometimes detailed in connection with a device communicating with at least one second device of the system 100, the method can also be implemented locally by a device, for example when it is used successively by several users exchanging asynchronous messages (like emails).

On the contrary, in some of its embodiments, the method can be implemented in a distributed manner between at least two devices 110, 130, 132, 134, 136, 140 and/or 150 of the system 100.

The term "module" or the term "component" or "element" of the device here means a hardware element, in particular wired, or a software element, or a combination of at least one hardware element and at least one software item. The method according to the invention can therefore be implemented in various ways, in particular in wired form and/or in software form.

There illustrates some embodiments of the method 300 of the present application. The method 300 can for example be implemented by the electronic device 200 illustrated in .

As illustrated in , the method 300 may comprise obtaining 310 contextual data from electronic messages exchanged between several individuals, for example by using at least one of the electronic devices of the system 100. It may for example be an access in reading the information contained in a file (known as a "log") keeping track of all the electronic messages exchanged within the group (emails in the case of the example detailed without limitation), or access to a knowledge base, or an on-the-fly collection of this data via a central server of a messaging tool, for example.

In the detailed embodiments, the data (or traces) obtained from an electronic message include, for example, a date and/or time (time stamp or "timestamp" according to the English terminology) of transmission and/or reception of the message, an identifier of at least one sender (or sender) and/or a receiver (or recipient) of the message, and/or a descriptive wording of the message (such as a title, or subject, of the message in the case of e-mail for example), or textual content of an e-mail or an instant message).

By individual, we mean an entity that can send or receive a message. It may therefore be a single person or a plurality of persons acting collectively within the framework of these exchanges, such as for example, in at least some of the detailed embodiments, a plurality of persons accessible through a same so-called alias email address or collectively using the same email address for sending a message. In the latter case, in certain embodiments, the associated contextual data will for example be attributed to a single “fictitious” individual associated with this email address and therefore with the plurality of people. In certain embodiments, the contextual data of an e-mail whose recipient is an alias e-mail address of a plurality of e-mail addresses may be modified to have as recipients the plurality of individuals identified by the plurality of e-mail addresses of the 'a.k.a.

We assimilate in this request the individuals in copy of an email to its recipients. Furthermore, in certain embodiments, to avoid for example duplicating certain information, the contextual data can be collected relative to messages sent only, or in other embodiments relative to messages received only. For reasons of simplicity, in the embodiments detailed in the present application, the contextual data is collected relative to messages sent only and therefore does not deal with receptions (effective or not of these messages), the terms recipients and receivers being used in this application without distinction to designate the same individuals.

For the following, it is also considered as a consequence that the timestamp of a contextual data relates to the timestamp of transmission of the corresponding electronic message.

In certain embodiments, the method 300 can comprise a preparation 320 (or preprocessing or “preprocessing” according to the English terminology) of the data obtained. For example, the preparation 320 can comprise filtering 321 (optional) of the data obtained, so as for example not to take into consideration contextual data considered as personal or confidential. This filtering 321 can take account, for example, of the descriptive wording of the message.

In the detailed embodiments, the filtering 321 can for example comprise a filtering of at least some of the contextual data according to their sender and/or their recipient(s). For example, the filtering can keep only the contextual data having a sender internal to the group and/or at least one recipient internal to the group. In certain embodiments, the contextual data having at least one internal recipient and at least one external recipient can be modified so as to retain in their recipient(s) only those internal to the group. As illustrated in , the method 300 can comprise a standardization 322 of the labels (or subject) of the emails to obtain standardized labels. The normalization can comprise for example a deletion, in labels, of prefixes or suffixes added over the transfers, and/or responses to the same message. It is a question, simply put, of obtaining in certain embodiments, for all the types of emails, labels that are closer (for example identical) to the label of a creation email relating to the same subject as these different types of emails.

For example, standardization 322 may include deletion of exchange words added automatically, during an email response and/or forwarding, to the wording (or subject) of the response email and/or forwarding ( or response and transfer) by a messaging tool that generated these emails, to characterize the email (e.g. "re", "fwd", "tr", etc.) and/or annotations present in the wording, such as annotations added for example manually by senders of e-mails (such as "URGENT" - "UPDATE" - "PLEASE READ", or initials), answering or forwarding a received e-mail, in the subject of the received e-mail). In certain embodiments, the deletion can thus be based on an algorithm such as the “Algorithm 1” algorithm presented below in the form of pseudocode, by way of a simple non-limiting example:

In this algorithm:

• AW designates the set of words (prefixes or suffixes for example) added automatically and/or manually;
• S denotes the set of email labels, the element corresponding to the subject (label) of the username email ;
• RS denotes the set of standard email labels, the element corresponding to the subject of the id email .

In certain embodiments, the normalized label can be stored, in association with the relevant contextual data (therefore in addition to the – non-normalized – label of the contextual data). In other embodiments, the normalized wording can be stored in replacement of the non-normalized wording in the contextual data, at least some of the deleted portions (such as the exchange words) also being stored in association with the contextual data concerned.

In some embodiments, the preparation may also include an enrichment 323 of the obtained data. So, in the example of the , the traces of the e-mails can be crossed, via the identifiers of the senders or receivers concerned for example, with other data internal or external to the group (for example data from Human Resources or from the organization of the group), such as an identifier of a hierarchical entity of membership and/or a profession of a sender and/or a receiver of an e-mail, so as to enrich the contextual data, or to associate them with these hierarchical entity identifiers and/or these professions . This enrichment 322 may be optional in certain embodiments.

In certain embodiments of the present application, the method may also comprise an (optional) anonymization 324 of the data obtained, so as for example to make them comply with a regulation (such as a national or international regulation). For example, in embodiments where the data obtained contains identifiers of the senders and receivers corresponding to their e-mail addresses, the method can comprise an optional step of anonymization comprising an application of a hash function to the addresses of emails from senders and/or receivers, so as to have at least one unique and anonymous identifier for each individual. Portions of the labels designating individuals (surname and first name for example) can in certain embodiments also be deleted.

Once the contextual data has been obtained, the method 300 can comprise obtaining 330 at least one communication chain (for example chaining contextual data) representative of the exchanges of at least one same piece of information (the content of the e-mails for example) between individuals in the group. In some embodiments, obtaining a communication chain can for example take into account at least some of the contextual data associated with the electronic messages conveying this information. For example, obtaining 330 a communication chain may include grouping 331 the contextual data taking into account a similarity between their standardized labels. It is possible, for example, to group together contextual data with the same standardized label, or very similar standardized labels, to take into account the absence of detection of certain prefixes or suffixes during the normalization and/or a probability of the presence of such prefixes or suffixes from standardization.

Note that the contextual data can come from electronic messages corresponding to different types of information exchange. Within the framework of the detailed embodiments, based on emails, one can for example distinguish at least 4 types of emails, and in particular as illustrated by FIGS. 4 to 7:

informational “creation” emails (also referred to herein as “zero emails”);

so-called “response” emails;

so-called “forwarding” emails;

so-called “reply and forward” emails.

The creation email is an email that creates at least one piece of information (the content of this email) without relying on a previous email, as shown in . Such an email is, in other words, an email that initiates a conversation involving individuals in the group. In the example shown in , a creation email is sent by individual A with individuals B, C and D as recipients.

Reply email is an email that responds to a prior email without involving any person(s) other than those already involved in the prior email, as shown in . In the example shown, A sends an e-mail to B at time instant (t -1), then at the later instant t, B sends a reply to this e-mail to people who are also recipients of A's e-mail. is newly added to the flow of information (in exchanges). A reminder email from a previous email can be considered a reply email.

The forwarding email is an email that forwards an earlier email by involving only person(s) other than those already involved in all of the emails in the chain prior to that forwarding, as shown in . In the example illustrated, A sends an e-mail to C at time (t -1), then at a later time t, C transfers this e-mail to persons (M and N) who are not recipients of A's e-mail. recipients of a transfer e-mail are individuals absent from the communication chain (unlike certain recipients of the reply and transfer e-mail introduced below).

The reply and transfer email is an email that responds to a previous email but also involves at least one person(s) other than the one(s) already receiving at least one email from the chain, prior to this transfer, such as illustrated in (thus effectively forwarding the email to that other person). In the example illustrated, A sends an e-mail to B at time instant (t -1), then at a later instant t, B sends a response to this e-mail by involving, in the receivers, on the one hand one or several people involved in the e-mail sent by A and on the other hand at least one person E. A return mail from a previous e-mail can be considered as a response and/or forwarding e-mail depending on its recipients.

We now present in conjunction with the a chain of communication.

For a given wording (for example a standardized wording), a set of emails, corresponding to different types of exchanges (as introduced above) is exchanged between individuals of the group (within a constant duration for example). This set (more precisely in the detailed embodiment the contextual data associated with these emails) forms a communication chain. There illustrates a communication chain, by representing it as a directed graph 810 whose nodes 811, 813 are individuals of the group involved in at least one email having this subject, a link 812 between two nodes representing a sending of an email by the individual of the origin node of the link, this email having among its recipients the individual of the destination node of the link. (For the sake of readability, only part of the links illustrated in has a numerical reference) The graph representing (or associated with) the chain comprises a root node (or origin) 813 which corresponds to the sender of the creation email of the chain, the other nodes 811 representing recipients of this creation email or reply, forward, or reply and forward emails based on this creation email (and therefore subsequent to this creation email). For example, in graph 810 of the , the chain is initiated by a creation email, sent by individual A to individuals C and to B, then a response and transfer email is sent by individual C to individuals A (response), D (transfer) , E (transfer) and F (transfer).

The recipients of a transfer email are all newly involved (or added), by this email, in the chain of communication introduced below. A reply-and-forward email newly involves (adds) at least one of its recipients in the communication chain.

For the sake of simplification (and unless otherwise indicated), we will assimilate transfer emails and response and transfer emails, for their transfer part. Similarly, response emails and response and forwarding emails will subsequently be assimilated (unless otherwise indicated), for their response part.

In some embodiments, as shown in , the obtaining (or construction) 330 of at least one communication chain, can comprise a scheduling 332 of at least one email grouped with the email for creating the chain (for example with the same standardized wording).

The positions of the nodes 811, 813 in the graph 810 associated with the chain are representative of these exchanges and their chronology. So in the nodes are represented with different levels. For example, sender A of e-mail zero is represented at a first "zero" level, recipients B and C of the creation e-mail being represented at the level immediately below (1 for example), recipients of the e-mail transferred by C being represented at level immediately below (2 for example) at the level of C. The respective recipients of the emails forwarded by D and E are represented at the level immediately below (3 for example) at the level of D and E. The node B which forwarded the email to C is represented at the same level as node C because chronologically it was added before this transfer (due to the sending of the creation email from A to B and C).

For example, scheduling 332 may include a search for at least one email identifier , chain of communication, previous emails ( ) and/or following ( ) chronologically immediately this email, as illustrated respectively by the algorithms “Algorithm 2” and “Algorithm 3”, below in the form of pseudocode, by way of simple non-limiting examples:

In “Algorithm 2”, the procedure identifies the email immediately preceding an email ID (resulting from preparation 320). The following notation is used (in addition to the notation already introduced in connection with the algorithm "Algorithm 1")

• denotes the email immediately preceding the username email ;
• is the set comprising the email, already identified, immediately preceding the identifier email . This set is initially empty.

All the contextual data obtained (step 310) and prepared (step 320) can for example be reviewed.

Different criteria can be used to determine whether an email ID whether or not precedes an email ID . For example, these criteria may include the following criteria:

• is different from ;
• The creation timestamp of the email ID is older than that of the username .
• Emails have the same or similar wording (when normalized).
• The sender of the email ID is a receiver of the email ID .

The email satisfying these criteria whose creation timestamp is closest to that of the username email , is considered the email immediately preceding the username email .

The following algorithm "Algorithm 2" can for example be implemented (the initial value corresponds in the algorithm to the notation of an empty set).

Similarly, the following “Algorithm 3” algorithm can for example be implemented. In the “Algorithm 3” algorithm, the procedure identifies the email created immediately after an email ID .

In the detailed embodiment, the scheduling 332 of a communication chain can take account, in addition to the chronology of emails of identical or similar standardized objects, of a chaining between sender(s) and receiver(s) of these e-mails, for example by applying chaining rules. For example, there may be chaining rules consistent with this chronology, such as the following rule: For any email created at time t other than the creation email, the sender of the email is the receiver of another email at a time t1 prior to time t.

For example, scheduling 332 may include a search for at least one email identifier , of the chain of communication, of all emails in the chain preceding and/or following this email (chronologically). According to a non-limiting example, presented below in the form of pseudocode by the algorithm "Algorithm 4", the procedure can allow to obtain (determine) a set of emails relative (or chained) to an email ID , based on the game .already introduced.

Such a search can for example be carried out for all the prepared contextual data.

As illustrated in , the obtaining 330 of at least one communication channel, may include an identification 333 of at least one creative email. This such email corresponds in the detailed example to an email that has no previous email Obtaining 330 of at least one communication chain can also include a construction 334 of a communication chain originating from the contextual data of an identified creation email and the other contextual data grouped together in the string being ordered with respect to the contextual data of this origin email in particular.

For example, in some embodiments, a string may be created for each email that has no previous email (ie, for each creation email). According to a non-limiting example, presented below by the algorithm "Algorithm 5", in the form of pseudocode, the procedure identifies all the creation emails of the obtained emails 320 and builds a communication chain for each of these creation emails using the algorithms presented above

In some embodiments, the method may include creating a communication channel for some of the creation emails only. For example, the method can filter certain e-mails (for example delete those associated with an object comprising the keyword "personal" or "confidential", or certain e-mails belonging to a "black" list of message senders and/or recipients. This filtering can for example be carried out in addition to the filtering 321 already presented above in connection with the preparation 320 of the contextual data.

In some embodiments, the method may include categorizing 340 the obtained communication strings. The categorization 340 can be performed in some embodiments based on information (or variables) descriptive of the communication channels, such as for example:

a number of individuals involved in the chain of communication email;

a number of individuals added to the communication chain after the creation email was sent (i.e. without taking into account the recipients of the creation email);

a total number of emails sent in the communication chain;

a percentage of creative emails in the communication chain, compared to this total number of emails sent;

a percentage of response emails in the communication chain, compared to this total number of emails sent;

a percentage of forwarding emails in the communication chain, out of this total number of emails sent.

In certain embodiments, the method may comprise obtaining 341 n-tuples each associated with a communication chain obtained 330 previously.

Each n-tuple can for example comprise an identifier c _i of a communication channel and values of descriptive variables of the communication channel. These values can be obtained for example in certain embodiments after applying algorithms as presented above.

These n-tuples can for example be stored in a database accessible to the device 200 of the .

In certain embodiments, the categorization 340 of the communication chains can comprise a partitioning 342 (clustering according to the English terminology) of the set of n-tuples obtained.

This partitioning can for example use the algorithm -means on the n-tuples obtained, to obtain clusters of tuples constructed as a function, for example, of values, or range of values, of certain descriptive variables of the communication channels.

According to a first example, we can rely on the graphical method called “Elbow” to choose the number of clusters. More precisely, the K-means algorithm can be implemented with different K values, a curve graphically representing for the different numbers of clusters K, the evolution of the variance between the clusters, so as to select the corresponding number of clusters graphically at a bend in the curve. Using this method can help select the number of clusters in a simple way.

According to another example, to find the number , we can vary the number of clusters (from 2 to 30 for example) and choose the one taking into account the Davies-Bouldin criterion. This criterion can indeed help to measure, for each cluster, a similarity index with the cluster which is closest to it. The value of k selected can for example be that which minimizes the mean, over all the clusters, of the similarity indices measured. The Davies-Bouldin criterion can help to obtain more reliable results than the so-called "Elbow" method (especially when the number of clusters considered increases) with low computational complexity, which can be very important when the contextual data set is very large. bulky.

For example, we can obtain 342 a partition based on values of variables such as the percentage of forwarding emails (which gives an indication of the number of people added after the creation email). Such partitioning can, for example, help to identify the clusters where the information spreads the most and therefore the communication chains whose identifiers are present in the tuples that make up these clusters.

In certain embodiments, the method may comprise a selection 343 of one or more clusters, among these k clusters. These clusters can for example be those which best correspond to target values of the descriptive variables (for example of those used for clustering). For example, in the illustrated embodiment, which focuses by way of example and for simplicity on the propagation of information, the method can comprise a selection 343 of the clusters obtained, to keep only clusters which contain have a percentage of non-zero transfer emails (and therefore corresponding to a non-empty propagation chain as presented below). This selection may be optional in some embodiments.

The present application is indeed detailed, by way of example, in embodiments helping to identify individuals having an influence on other individuals (inside or outside the group of individuals). In particular, at least certain embodiments can help to identify at least one influential individual (or influencer) within a group such as a company and this in a more objective manner than certain solutions of the prior art. In addition, this application is concerned with the propagation of information. For example, an influencer can initiate a conversation and with the reaction of other individuals, the information can spread. This notion of propagation is different from the centrality of individuals in the exchanges (for example, all conversations go through an individual), assimilated by certain solutions of the prior art to the importance of the people involved in the exchanges. Indeed, this notion of importance (often linked only to a hierarchical position) differs from the notion of influence, which does not prejudge a person's hierarchical position in exchanges. Moreover, when it comes to a chain of communication corresponding to a simple exchange between two individuals, not provoking a response or transfer message, it is very difficult to speak of influence or to know whether these two individuals (or one of them) are influencers or not.

The authors considered that one of the main characteristics of an influencer was his ability to make another individual react. In terms of dissemination of information by messages, this reaction results in a propensity to push another individual to disseminate in turn the information coming from the influencer, that is to say in the detailed embodiments, forward the email received from an influencer. For this reason, the present application proposes to search for influencers taking into account the propagation of emails in communication chains.

In the embodiments detailed in the present application, the descriptive variables used for the clustering can include in particular variables making it possible to characterize a propagation of information within a communication chain.

The communication chains obtained (step 330) can involve a very variable number of individuals, ranging from 2 individuals (for example sending an email followed by a response), to a large number of individuals, for example a few hundred of individuals for a group of several thousands or tens of thousands of individuals (representative for example of mass mailings to a very large number of individuals, and/or of significant propagation via numerous transfer emails). This variable number of individuals is representative of a variable propagation (or dissemination) of the information contained in the creative email at the origin of the communication chain).

The present application proposes, in at least some of its embodiments, to help characterize the propagation of information of a creative email.

The process 300 illustrated in may comprise an analysis of at least one communication chain representative of a propagation (also called a propagation chain). According to certain embodiments, this string can for example be chosen from a cluster obtained during the categorization 340 (for example a cluster selected 343 previously). For example, the method can comprise a selection 343 of at least one of the “k” clusters obtained according to the method of the present application, selection being made according to variables used by the partitioning and representative, alone or in combination, of a propagation of an e-mail among several individuals. For example, the method can select at least one of the clusters of chains by taking into account variables such as in particular a percentage and/or a number of forwarding emails in at least one chain of a cluster (compared to a number of total emails in the chain) and or a number of individuals involved in the chains of the cluster (therefore a depth of the associated graph). For example, a propagation chain can belong to a cluster comprising chains having a non-zero percentage of forwarding emails.

A propagation chain is a communication chain (built around a creation email) representative of a propagation (or dissemination) of the information contained in this creation email (for example its subject and/or its content) by individuals over time. Within a communication chain, the dissemination of information is carried out (prior to the creation email) thanks to emails that add at least one new individual to the chain, i.e. the various transfer emails .

For example, in , A receives an e-mail, forwards it to B and C, then B reacts to A's e-mail by in turn forwarding the e-mail to D (which he thus adds to the chain of communication), thus propagating the information initially transmitted by AT.

In at least some of the detailed embodiments (where it is sought, by way of illustrative example, to characterize a propagation of information via at least one communication chain), the method of the present application may comprise obtaining 350 of at least one transfer graph, a transfer graph being obtainable 350 from the directed graph associated with a propagation chain of a selected cluster 343. Obtaining 350 the transfer graph can comprise a deletion of this graph oriented links corresponding to a response email, to keep in the transfer graph only the links representative of a transfer to an individual not involved in the emails, prior to the transfer, exchanged from the creation email. In other words, in certain embodiments, only the links are kept in the transfer graph which correspond, in view of the temporal ordering of the nodes of the directed graph, to additions of at least one new node to the directed graph.

In such an embodiment, the transfer graph can therefore contain the set of nodes added over time as well as the set of links representing the addition of a new node in the directed graph. . There represents, by way of example, the transfer graph 820 constructed from the directed graph 810 already commented on above.

In certain embodiments, the method 300 can comprise obtaining 360 (calculation) of an indicator called degree of influence (or measure or score of influence), relating to at least one individual belonging to at least one transfer graph , and representative of the reactions of the individuals of the transfer graph to which the individual belongs, after the sending of an email by this individual. This reaction is represented by the addition of new people in this transfer graph, as already explained (in a transfer email).

In certain embodiments, the method 300 can comprise an association of a weight with at least one link of at least one sub-graph of a propagation graph (in other words a weighting of this link). The subgraph of a node of a transfer graph is called the part of the transfer graph associated with the propagation chain having this node as its origin and therefore comprising this node and the nodes added by this node to the transfer graph. The weight of a link can for example be different depending on whether the link relates to a receiver who has himself reacted by adding at least one new individual or not. For example, in certain embodiments a higher weight can be assigned to the links including the receiver which has itself reacted by adding at least one new individual. In the transfer graph 820 presented in figures 8 and 9, we see for example that individual C added 3 people (D, E and F) and was able to make only 2 people of these 3 people react (this is i.e. D and E, which added G and H respectively). In the example of the , the link from C to D and to the link from C to E are assigned a strictly positive weight 822 such as the value 1), the link from C to F on the other hand receiving a weight 822 of zero. The degree of influence of a node can be calculated in certain embodiments from the weights of the links of which it is the origin (for example as a sum of these weights).

In some embodiments, to measure the degree of influence of an individual in a propagation chain, it is possible, for example, to use the following formula:

Or

) is the individual's measure of influence in the transfer graph representing the propagation chain .

is the number of individuals added by In

is the number of individuals added by and who reacted.

is the maximum number of people added in the transfer graph representing the chain , i.e. the upper limit, over the set of nodes of the transfer graph, of the number of individuals added by these nodes, therefore the maximum, over all the a of C, of the with the notation above;

the number of nodes of the subgraph of the individual in the transfer graph representing the propagation chain

The influence measure can be calculated in the subgraph associated with at least one individual of at least one transfer graph of at least one propagation chain (for example with all the individuals of all the transfer graphs obtained) .

In the example above, the measure is between 0 and 1. It is 0 when the individual does not add other individuals or when none of the added individuals reacted. It tends to 1 when the individual has added a large number of individuals in the transfer graph (relative to the maximum number of individuals at the size of the chain) and all these individuals have reacted.

We now place ourselves as an example in the use case where two individuals have succeeded in making the same number of individuals react in a given chain. The first individual has reacted only individuals of the same entity as him and they work together daily (they are very close in terms of communication). The second individual has caused reactions from individuals who do not work with him on a daily basis (they are distant in terms of communication, for example). In addition, these individuals work in different and/or external entities.

In some embodiments, the second individual may be considered to have a greater degree of influence than the first individual.

However, the measure proposed above does not make it possible to differentiate the two individuals in terms of influence.

Also, in certain embodiments, for example embodiments where the preparation 320 of the contextual data includes an enrichment 322 of these data, the measure of influence proposed above can be modified by adjusting the variable so as to take into consideration a proximity between at least two individuals in terms of communication and/or a hierarchical proximity between an individual and an entity of the group.

The modified measure can take, for example, the following form:

Or

is the set of individuals added by in the transfer graph associated with the chain and who reacted by adding at least one new individual.

is the proximity between the individual and the individual . This measure takes the following form for example:

Or

is the number of emails sent from towards in a fixed or chosen duration (for example a constant duration).

is the set of individuals to whom sent at least one email in the time period considered.

For example, in the example detailed here without limitation, the duration may correspond to a sliding time window of constant duration, such as a duration of the order of ten days, or of the order of one or a few months (for example 6 months)

The duration can for example be chosen so that the entities of the group have a relatively stable composition (with individuals keeping the same functions and attached entity during this duration for example). The duration may for example vary according to the size of the group and/or the stability of the entities within the group.

• is the proximity between the individual and the entity of the individual . This measurement can for example take into account the hierarchical structure of the group, an entity being able, in certain embodiments, to be closer to its subordinate entities than to another entity which is not subordinate to it.

In certain embodiments, obtaining the measure of influence may depend on belonging to at least one zone of a plurality of zones (or groups) of entities defined according to hierarchical proximity. For example, in some embodiments, the following two areas may be defined:

A first zone, said to be specific to an individual rated containing the entity of and its sub-entities (i.e. entities directly or indirectly attached to the entity of ). For example, if the individual belongs to the entity ET1/ET2 then the proper space of this individual is the set of entities {ET1/ET2, ET1/ET2/ETA, ET1/ET2/ETB, etc} .

A second zone, called external to an individual rated which represents the set of entities that do not belong to the first zone specific to , and includes the outside of the group.

In some embodiments, reacting an individual who belongs to the own zone may be considered less important than reacting an individual belonging to the outer zone. The formula of below can for example be proposed:

Or represents the proximity of with its own area.

can for example be defined by the following formula:

Or is the number of entities belonging to the eigenzone of

In the example above, the measure e is between 0 and 1. It is worth 0 when the individual does not add a new individual to the chain or when the individuals added by this individual have not reacted to his email. This measure will be close to 1 if the individual adds the maximum number of individuals in the chain, if all the added individuals participate themselves in making the chain grow and if the added individuals have both an "empProx" individual proximity and a proximity entity "entProx" close to 1.

In the detailed embodiment, the method may comprise obtaining 370 from at least one propagation chain, at least one identifier of at least one "influencer" individual (for example by analysis and modeling of profiles influencers) taking into account, for example, one or more target characteristics such as:

a transfer depth, i.e. a number of successive transfers in the chain;

a number of emails sent by an individual in terms of the time period considered;

a fluctuation in the number of emails sent by an individual over the period considered;

the recipients of the emails sent by the individual over the period in question;

a number of recipients of emails sent by the individual over the period considered

a fluctuation of the recipients of the emails sent by the individual over the period considered

a fluctuation in the number of recipients of emails sent by the individual over the period considered

a location in the group of the recipients of the emails sent by the individual over the period considered (for example their membership of entities close or far from the entity of the individual)

a frequency of sending, by the individual and over the period considered, e-mails to the same recipient;

a frequency of sending, by the individual and over the period considered, emails creating chains of propagation;

a frequency of sending, by the individual and over the duration considered, of emails creating chains of comprising a number of nodes greater than a first value, such as or a number of nodes greater than a first threshold value.

In the detailed embodiment, the obtaining 370 of at least one influencer individual identifier can for example take into account the degree of influence (obtained 360 as described above for example) of at least one individual of at least one propagation chain (for example of the degree of influence of each individual of each propagation chain).

In certain embodiments, it is possible to take into account all the degrees of influence obtained by an individual for all the propagation chains to which he belongs. To do this, it is necessary, for example, to evaluate the behavior of individuals on the whole of the data.

In certain embodiments, the obtaining 370 of the identifier of an individual influencer can for example take account of at least one other criterion. For example, in certain embodiments, the method can take into account a depth of at least one chain to which an individual belongs (this depth possibly being representative, in certain situations, of a propagation of information). In some embodiments, the method may include comparing an individual's degree of influence in a chain against a first numeric value. This value can for example correspond to a threshold value of the degree of influence of an individual, whatever the location of the individual in the chain considered (beginning or in the middle of the chain), or in all of the channels to which it belongs, so that it is considered to have influence. In one embodiment, as in the detailed embodiments, where the influence is considered to be greater the greater the influence measure, it may for example be a minimum value.

In certain embodiments, the obtaining 370 of the identifier of an individual influencer can take into account, for example, the relationships between at least two individuals of a chain. Indeed, there can sometimes be functional relationships between individuals within the group which can induce a large number of email exchanges between these individuals. For example, an individual whose role is to assist another individual will generally react after receiving an e-mail from this other individual (according to the instructions of this other individual for example) and will overwhelmingly receive e-mails from this other individual.

In certain embodiments, the method can take into account a number of senders of the emails of which an individual is the recipient (if he is located in the middle of the chain). These issuers are subsequently called “pre-influencers”.

In certain embodiments, to identify for example individuals having a great influence within a group, the method can comprise a partition (for example via the algorithm known as the name K-means) of a set of data descriptions of the individuals in the group.

For example, the method can comprise an association with an individual of at least one descriptive variable, such as for example a descriptive variable taking into account at least one of the following elements:

a number of chains where the individual is at the origin of the chain (initiation of the chain);

a depth of at least a portion of at least one transfer graph, the portion having the individual as its origin;

a number of chains for which the portion of the transfer graph having the individual as its origin has a depth greater than or equal to a first depth value;

a number of email senders of which the individual is a receiver (ie a number of pre-influencers);

a number of chains in which the individual obtained the highest measure of influence in the chain;

a number of chains in which the individual obtained the second highest influence measure of the chain;

a number of chains in which the individual obtained the third highest influence measure in the chain; and or

a number of chains in which the individual obtained one of the N highest measures of influence in the chain; with N strictly positive integer, varying according to the embodiments.

The choice of the number ( ) of clusters, can for example be carried out by varying the number of clusters (from 2 to 30 for example) and by choosing, from among the varied numbers of clusters, a number of clusters taking into account the Davies-Bouldin criterion. Individuals of influence can for example be determined according to their membership (or non-membership) of certain clusters

The present application, in at least some of its embodiments implementing an automatic analysis of the exchanges between individuals, help to propose an objective method of analysis, based on actual exchanges between individuals, unlike certain methods of the prior art, based on manual feedback (such as ratings) by certain users, and therefore on subjective assessments and potentially representative of only a small number of individuals.

In some embodiments, the method may further comprise storing information obtained by the method of the present application, such as influence measurements obtained and/or at least certain descriptive values, in association for example with at least an individual, so as for example to enrich a knowledge base relating to at least some of the senders and/or receivers of the messages. In certain embodiments, in particular when the identifier of the sender and/or of the receiver of the message has not previously been "anonymized", or can be identified (for example "de-anonymized" via a correspondence table linking identifiers anonymous or not from the same senders), this information can be associated with the sender or the receiver of the message (for example, they can be added to his personal data, so as to enrich his "profile"). As a variant, this information can be associated as a variant with at least one terminal of the sender or of the receiver of the message.

Alternatively or cumulatively, in certain embodiments, this information can be associated with an entity of several individuals to which the sender and/or at least one receiver of the message belongs (or as a variant an entity associated with several terminals of which at least least one belongs to the sender and/or to a receiver of the message), for example an entity identified, during an anonymization 321 of an identifier of the sender and/or of the receiver of the message as explained below.

The method of the present application can help to obtain results associated with individuals or entities, depending on at least one descriptive value sought (targeted for example for the subsequent exploitation of this analysis). These descriptive values can for example be used within the framework of the partitionings described above, for the constitution of the clusters and/or the selection of at least one of the clusters for example. Thus, when the method is applied for a group of employees of a company, it may be a set of descriptive values making up a standard profile sought (or on the contrary to be avoided), such as for example a representative standard profile influence within the group. The targeted descriptive values can be obtained in various ways depending on the embodiments. For example, they can be obtained at least partially via a communication interface or a user interface of the device 200 of the , and/or from a storage means accessible from the device 200 (for example of one or more configuration file(s), which can be local to the device or remote).

The method may include an optional step of rendering (or restitution), textual, visual and/or audio, of the results of at least some of the steps of the method of the present application, on a man-machine interface of the device 200 or via another device coupled to device 200 (printer, electronic data storage medium, etc.)

At least certain embodiments described above can also be used in a tool for steering the strategy of a company.

At least certain embodiments described above can also be implemented by devices used by communication members of a company, or by team managers, or persons in charge of deploying various tools or methods in a business.

As an alternative to the embodiments detailed above, one could use descriptive values taking into account or relating to subjects of interest (or themes) in the objects and/or content of the messages, so as to identify influencers by theme.

According to another variant, one could use descriptive values taking into account or relating to an individual's propensity to react to a message, so as for example to help predict at least partially the reactions (or lack of reaction) of at least one individual has an email.

Claims (10)

Method implemented by at least one communication device, comprising:
obtaining a plurality of contextual data relating to electronic messages transmitted or received by a plurality of terminals of a communication network, one contextual data of said plurality of contextual data comprising at least:
- a description of an electronic message;
- At least one identifier of at least one sender and at least one receiver of said message; And
- a timestamp of said message;
obtaining at least one first contextual data item from said plurality of contextual data items or at least one identifier of a first sender of at least one first contextual data item from said plurality of contextual data items, depending on a similarity between labels of said contextual data and of a similarity between values of at least one descriptive information of said contextual data and at least a first value of said descriptive information. Method according to claim 1, in which said obtaining of said identifier of said first sender and/or of said first contextual data comprises grouping said contextual data according to a similarity between their labels. Method according to claim 2, in which said obtaining of said identifier of said first transmitter and/or of said first contextual data item comprises chaining at least a portion of at least two of said grouped contextual data, said chaining taking into account the identifiers of the transmitters and /or receivers and/or timestamps of said grouped contextual data. Method according to claim 3, in which said chaining takes account of portions of said labels indicative of a propagation, between at least two of said transmitters and/or receivers of said grouped contextual data, of said electronic messages to which said grouped contextual data relate. A method according to claim 4 wherein said obtaining said identifier of said first sender and/or said first contextual data comprises normalizing the labels of said contextual data producing normalized labels, said normalizing comprising deleting said portions indicative of said labels from said contextual data, said grouping taking into account said standardized wordings. Method according to any one of Claims 2 to 5, in which the said descriptive information relates to the said contextual data grouped by the said grouping. Method according to any one of Claims 2 to 6, in which the said descriptive information is representative of a propagation of an electronic message to which at least one second of the said grouped contextual data relates, between the said transmitters and/or receivers of the said grouped contextual data. with said second contextual datum. Method according to claim 7 wherein the timestamp of said second contextual data is earlier than the timestamps of said contextual data grouped with said second contextual data Method according to Claim 8, in which the said descriptive information belongs to a group comprising:
a number of transmitters and/or receivers of said second contextual datum;
a number of senders and/or receivers other than the senders and/or receivers of said second contextual datum in said at least one of said contextual data grouped with said second contextual datum;
a total number of contextual data grouped with said second contextual data;
a percentage of contextual data with the same label and the same timestamp as said second contextual data, in said at least one of said contextual data grouped with said second contextual data, relative to said total number;
a percentage of contextual data relating to a response message to another message, in said at least one of said contextual data grouped with said second contextual data, relative to said total number;
a percentage of contextual data relating to a transfer message of another message, in said at least one of said contextual data grouped with said second contextual data, relative to said total number. Electronic device comprising at least one processor, said processor being configured to implement the method according to one of Claims 1 to 9.