EP4256444A1

EP4256444A1 - Digital data tracing method

Info

Publication number: EP4256444A1
Application number: EP21823871.5A
Authority: EP
Inventors: Charlotte JACOBÉ DE NAUROIS; Claire Laudy
Original assignee: Thales SA
Current assignee: Thales SA
Priority date: 2020-12-03
Filing date: 2021-12-01
Publication date: 2023-10-11
Also published as: WO2022117677A1; FR3117228A1; US20240020341A1

Abstract

The present invention relates to a digital method for tracing data originating from a plurality of data sources, the method comprising the steps of: a. obtaining initial data originating from a plurality of data sources, each initial datum comprising initial elements and initial connections, each initial element having a type and a value; b. determining a final datum on the basis of at least two initial data following the merging with one another of initial elements of the initial data according to a merging strategy, the merged initial elements making it possible to obtain a merged element having a type and a value; and, c. determining a merging history for each merged element, the merging history comprising the type and the value of each initial element from which the merged element originates, the type and the value of the merged element and data relating to the merging strategy.

Description

TITLE: Digital data tracing process

The present invention relates to a method for digital plotting of data from several data sources. The present invention also relates to a computer program product and an associated readable information medium.

In areas such as surveillance, a lot of data from different data sources is collected and analyzed.

To reduce redundancies and thus facilitate data analysis, data compression techniques, also called data fusion, are used.

However, once the merger is complete, it is difficult to trace the original information. It is also not easy to understand and verify the merging criteria.

However, this lack of monitoring of the operations carried out is problematic for certain critical applications, in particular when certain data sources are considered less reliable than others. This makes the merged data less reliable and more difficult to exploit.

There is therefore a need for a method making it possible to improve the understanding of a fusion of data with a view to facilitating the exploitation of the merged data.

To this end, the present description relates to a method for digital plotting of data from several data sources, the method being implemented by computer and comprising the steps of: a. obtaining initial data from several data sources, each initial data comprising initial elements and initial links, each initial element having a type and a value, each initial element belonging to a first category of elements, called initial entities, or to a second category of elements, called initial relations, the initial links connecting initial entities to one another via an initial relation, b. determination of a final datum as a function of at least two initial datum following the fusion of initial elements of said initial data with each other according to a fusion strategy, the merged initial elements belonging to the same category of elements and making it possible to obtaining a merged element, the merged element having a type and a value, and c. determining a merge history for each merged item, the merge history including the type and value of each initial item from which the merged item originated, the type and value of the merged item, and data about the merge strategy.

According to particular embodiments, taken in isolation or according to any technically possible combinations, the method comprises:

- the merge strategy implements functions, each function being represented by a set of function elements each comprising a type and a value, the merge history comprising history elements and history bindings, each history element having a type and a value, the historical elements belonging to: a. a first category of historical elements called historical entities, grouping together the initial elements, the merged element and the function elements representative of the functions, and b. a second category of historical elements, called historical relations, historical links linking historical entities to each other through historical relations;

- historical relations and historical links are representative of the merging strategy applied to the initial elements via the function elements to obtain the merged element;

- each initial element further has an initial history comprising the type and the value of the initial element, as well as data identifying the data source from which the initial element originated, the determined merge history for each merged item including the initial history of each initial item from which the merged item originated;

- each historical entity is associated with at least one additional piece of data chosen from: c. a timestamp indicating the date and/or time of generation of the historical entity, d. data relating to the place of generation of the historical entity, e. data relating to the computer by which the historical entity was generated, and f. data relating to a user of the computer by which the historical entity was generated;

- the initial data of at least one data source come from measurements made by a sensor;

- the process comprises: g. a step of acquiring a determined merge history display command for a merged element, and h. a step of displaying the fusion history of said merged element on a display;

- the initial data relate to at least one piece of equipment, the method comprising a step of implementing an action on the piece of equipment according to the final data(s) obtained, and at least one fusion history.

The present description also relates to a computer program product comprising a readable information medium, on which is stored a computer program comprising program instructions, the computer program being loadable on a data processing unit and implementing a method as previously described when the computer program is implemented on the data processing unit.

This description also relates to a readable information medium on which a computer program product as described above is stored.

Other characteristics and advantages of the invention will appear on reading the following description of embodiments of the invention, given by way of example only and with reference to the drawings which are:

Figure 1, a schematic view of an example of a computer allowing the implementation of a digital data plotting process,

Figure 2, a flowchart of an example implementation of a digital data plotting process,

Figure 3, a schematic representation of an example of a final data G obtained following the fusion of two initial elements G1 and G2 of two distinct initial data,

Figure 4, a schematic representation of an example of three initial histories, each associated with a distinct initial element,

Figure 5, a schematic representation of another initial history example,

Figure 6, a schematic representation of an example of a merge history associated with a merged element obtained following two merges, the merge history having been represented in two parts for the sake of readability, Figure 6 illustrating only the first part,

Figure 7, a schematic representation of the second part of the merge history shown in Figure 6, and

Figure 8, a schematic representation of different entity types and subtypes. A computer 10 and a computer program product 12 are illustrated in Figure 1.

Computer 10 is preferably a computer.

More generally, the computer 10 is an electronic computer suitable for manipulating and/or transforming data represented as electronic or physical quantities in computer registers 10 and/or memories into other similar data corresponding to physical data in memories, registers or other types of display, transmission or storage devices.

The computer 10 interacts with the computer program product 12.

In the example illustrated by FIG. 1, the computer 10 comprises a processor 14 comprising a data processing unit 16, memories 18 and an information carrier reader 20. In the example illustrated by Figure 1, the computer 10 also includes a man-machine interface 22, such as a keyboard, and a display 24.

The computer program product 12 has an information carrier 26.

The information medium 26 is a medium readable by the computer 10, usually by the data processing unit 16. The readable information medium 26 is a medium suitable for storing electronic instructions and capable of being coupled to a computer system bus.

By way of example, the information medium 26 is a USB key, a floppy disk or floppy disk (from the English name “floppy disk”), an optical disk, a CD-ROM, a magneto-optical disk, a memory ROM, RAM memory, EPROM memory, EEPROM memory, magnetic card or optical card.

On the information carrier 26 is stored the computer program 12 comprising program instructions.

The computer program 12 can be loaded onto the data processing unit 16 and is adapted to cause the implementation of a digital data tracing method, described in the following description, when the computer program 12 is implemented on the processing unit 16 of the computer 10

The operation of the computer 10 will now be described with reference to FIG. 2, which schematically illustrates an example of implementation of a digital data plotting method and to FIGS. 3 to 8 which are examples making it possible to illustrate certain steps of the process.

By the term “digital tracing of data”, it is understood the fact of associating information with a datum indicating the origin of the datum and/or the actions having made it possible to obtain the datum from other data. In the description, the terms "melting" and "compression" are used equivalently.

In the example described below, the digital plotting method is implemented by the computer 10 in interaction with the computer program product 12, that is to say is implemented by computer.

The plotting method includes a step 100 of obtaining initial data Di. The initial data Di typically comes from several distinct and heterogeneous data sources. By the term "heterogeneous", it is understood that the same information can be described differently depending on the data source considered.

The initial data Di are, for example, data relating to individuals, installations or equipment.

Data sources are databases of information. For example, the data sources are digital libraries, social networks, blogs, files created by an operator or a machine, GPS signal traces, or even results of sensor data analysis (radar, AIS for example). Advantageously, the initial data Di from at least one data source come from measurements made by a sensor. The sensor is, for example, a camera, a sound sensor, or any sensor for measuring a physical quantity (radar, AIS for example).

Each initial datum Di comprises initial elements Ei and initial links Li.

Each initial element Ei has a type and a value. The type of an element corresponds to a concept represented by the element. Each type is associated with a set of possible values. The value of an element corresponds to the value taken by the type of the element among the set of possible values.

Each initial element Ei belongs to a first category of elements, called initial entities Ci, or to a second category of elements, called initial relations Ri. The initial elements Ei belonging to the first category of elements differ from those of the second category of elements by the types of said elements which are different.

Initial links Li connect initial entities Ci to each other via an initial relation Ri. In particular, the initial entities Ci are representative of states (identity of a person, place) and the initial relations Ri establish a link between these states (action of one entity on another).

In one embodiment, each initial datum Di is thus in the form of an ontological representation. The ontological representation is, for example, a conceptual graph or a logical representation of order one, that is to say a representation comprising alphanumeric symbols and symbols of punctuation. The symbols represent relationships between objects in the real world (Example: location ([Car: “Titine”@V1], [Position: “48.853, 2.35”@ND] expresses that the car V1 is called Titine, and is positioned at the ND location which is located at GPS coordinates 48.853; 2.35).

In general, an ontology is an entity formed by a set of knowledge, facts and rules relating to a given domain, for example, scientific, cultural, administrative, industrial or commercial know-how.

More precisely, an ontology is a set of concepts, each concept being able to relate to other concepts. A concept brings together a set of words designating the same notion, typically synonyms. For example, an aircraft is a concept that can also be referred to as a flying machine. However, the case of synonyms is not the only possible case. For example, for the aircraft concept and the concept the assets of a company, a new concept can be the aircraft of the said company.

An ontology can be represented in the form of a graph whose points are the concepts, a relation being represented by a line between two points, as is the case for conceptual graphs. An ontology can also be represented in the form of alphanumeric symbols, as is the case for first-order logical representations.

In the example illustrated by FIG. 3, two initial data items Di are illustrated, one in the form of a conceptual graph G1 and the other in the form of a conceptual graph G2. In this example, each initial datum Di comprises two initial entities Ci, a relation Ri and two links Li connecting each initial entity Ci to the relation Ri. In particular, the initial data Di represented by the conceptual graph G1 comprises: a first initial entity Ci whose type is “Man” and the value “John Dow”, a second initial entity Ci whose type is “Country” and the value “Philippines”, and a relation Ri whose type is “Relation” and value “Lives”. The initial data Di represented by the conceptual graph G2 comprises: a first initial entity Ci whose type is "Person" and the value "J. Dow", a second initial entity Ci whose type is "Country" and the value "Africa from the south", and an Ri relationship whose type is "relation" and value is "Travel".

Those skilled in the art will understand that the representation of the data in the form of conceptual graphs is given by way of examples, these data being able to be represented in other forms. For example, a first-order logical representation of the conceptual graph G1 is: Lives ([Man: John Dow], [Country: Philippines]).

Advantageously, each initial element Ei is, moreover, associated with an initial history Hi allowing the traceability of the initial element Ei, that is to say to go back to information relating to the origin of the initial element Ei. The initial history Hi thus comprises the type and the value of the initial element Ei, as well as data identifying the data source from which the initial element Ei originated.

In an example implementation, the initial history Hi has the same form as the initial data Di. Thus, if the initial data Di is in the form of an ontological representation, such as a conceptual graph, so will the initial history Hi. In this example, the initial history Hi includes initial history elements Ehi and initial history links Lhi. Each initial Ehi history element has a type and a value. Each initial historical element Ehi belongs: either to a first category of initial historical elements called initial historical entities Chi, or to a second category of initial historical elements, called initial historical relations Rhi.

The initial historical links Lhi connect initial historical entities Chi to each other through the intermediary of the initial historical relations Rhi. In particular, the initial historical entities Ci are representative of the initial elements Ei and of the data sources, and the initial relations Ri establish a link between these elements.

Advantageously, the initial historical relations Rhi and the initial historical links Lhi make it possible to establish the origin of the corresponding initial element Ei.

In the example illustrated by figure 4, three examples of initial histories Hi1, Hi2, Hi3 are represented. Each initial history Hi1, Hi2, Hi3 presents a first initial historical entity Chi representative of the initial element Ei to which said initial history Hi corresponds, a second initial historical entity Chi representative of the data source from which the initial element Ei originates, and an initial historical relation Rhi establishing a link between the initial element Ei and the data source from which it originated. Thus, for the first initial history Hi1 which is the history of the initial entity “Man: John Dow”:

- the first initial historical entity has the type "Man" and the value "John Dow" (corresponds to the initial element),

- the second initial historical entity has the type "Source" and the value "Source A", and

- the initial historical relationship has the type “relation” and the value “A resource”. Advantageously, each initial historical entity Chi of an initial history Hi is associated with at least one timestamp indicating the date and/or time of generation of the historical entity Ch. In the example illustrated by FIG. 4, the The timestamp is shown as a succession of numbers in each initial historical entity Chi. This timestamp is, for example, of the UNIX type, ie it designates the number of seconds elapsed since January 1, 1970 at midnight UTC.

Advantageously, each initial history Hi comprises at least one initial history entity Chi chosen from an entity representative of: a date and time of creation of a data source (timestamp), a file considered in a data source, and information unit represented on one or more indexed lines of a file.

Thus, in the example illustrated by figure 5, which is an extension of the first initial history Hi 1 illustrated in figure 4, three additional initial historical entities Chi and three additional initial historical relations Rhi are represented, namely:

- an initial historical entity Chi whose type is "Date-time" and whose value is "11/03/2020 14:20",

- an initial historical entity Chi whose type is "File" and whose value is "C/Documents/aaa.xls",

- an initial historical entity Chi whose type is "Line" and whose value is "1446",

- an initial historical relationship Rhi establishing a relationship between the initial historical entities "Source: Source A" and "Date-time: 11/03/2020 14:20", whose type is "Relation" and the value "A date- hour ",

- an initial historical relationship Rhi establishing a relationship between the initial historical entities "Source: Source A" and "File: C/Documents/aaa.xls", whose type is "Relation" and value "A file", and

- an initial historical relationship Rhi establishing a relationship between the initial historical entities "File: C/Documents/aaa.xls" and "Line: 1446", whose type is "Relation" and value "A line".

Alternatively, the initial history Hi of each or at least one entity is empty. This is particularly the case when the original data source is unknown.

The plotting method includes a step 110 of determining at least one final datum Df. The final data Df comes from the fusion of at least two initial elements Ei belonging to different initial data Di, in particular from different data sources. The merge is performed according to a merge strategy.

At the end of the fusion, a merged element Ec having a type and a value is obtained. For example, the type of the merged element Ec is the type of one of the initial elements Ei from which the merged element Ec originated; and the value of the merged element Ec is the value of one of the initial elements Ei from which the merged element Ec originated. The merging strategy stipulates in particular that the merged elements belong to the same category of elements, that is to say either all to the first category of elements, or all to the second category of elements.

In particular, in one example, the merging strategy implements functions F, namely:

- a specific compatibility function to test if the types of the elements to merge are compatible,

- a similarity function specific to testing the similarity between the values of the elements to be merged according to a similarity threshold, and a merging function specific to defining the type and the value of the merged element Ec resulting from the merger.

Each function F has, for example, a name, input data, one or more parameters and output data (results). This information (input data, parameters, output data, name of the function) is, for example, represented by at least one type and at least one corresponding value in a history (merge history), each type and each value associated defining an element, called a function element.

The fusion is, for example, carried out according to a method as described in document FR 2 986 636 A, document FR 2 940 487 A or in the article by Laudy, C., Ganascia, J. G., & Sedogbo, C (2007, July), High-level fusion based on conceptual graphs. In 2007 10th International Conference on Information Fusion (pp. 1 -8) IEEE.

An example of final data Df in the form of a conceptual graph G is illustrated in figure 3. The final data Df comes from the initial data Di represented by the conceptual graphs G1 and G2. In particular, the entity Ci “Man: John Dow” of the initial data Di (first data) represented by the first conceptual graph G1 has been merged with the entity Ci “Person: J. Dow” of the initial data Di ( second datum) represented by the second conceptual graph G2. The result of the merge is the merged entity This "Man: John Dow". So in this example, the type and value of the merged entity is the type and value of the first data. Apart from the initial merged entities Ci which have been replaced by the merged entity Ce, the final data Df includes the same elements and links as the initial data Di.

The tracing method includes a step 120 of determining a fusion history Hc for each merged element Ec.

The fusion history Hc is a datum making it possible in particular to go back to the initial elements Ei from which the merged element Ec originated. The fusion history Hc thus includes information relating to the type and to the value of each initial element Ei from which the merged element Ec originated, the type and value of the merged element Ec and the data relating to the merging strategy.

In an example implementation, the fusion history Hc has the same form as the initial data Di and the final data Df. Thus, if the initial data Di is in the form of an ontological representation, such as a conceptual graph, it will be the same for the final data Df and the fusion history Hc.

In this example, the Hc merge history includes Eh-history elements and Lh-history links. Each historical Eh element has a type and a value. The types of the historical elements Eh notably include the types of the initial elements Ei, of the merged element Ec and of the function elements representative of the functions F used during the implementation of the merging strategy. The values of the entities of the historical graphs are for example the values of the initial elements Ei, of the merged element Ec and of the function elements representative of the functions F used during the implementation of the merging strategy. For function elements, the values represent, for example, the name of the chosen compatibility function, the name of the chosen similarity function, the name of the fusion function or the value of the similarity threshold.

The historical elements Eh belong: either to a first category of historical elements called historical entities Ch, grouping together the initial elements Ei, the merged element Ec and the function elements, or to a second category of historical elements, called historical relations Rh.

The historical links Lh connect historical entities Ch between them via historical relations Rh. In particular, the historical entities Ch are representative of the initial elements Ei, of the merged element Ec and of the function elements representative of the functions F of the fusion strategy, and the initial relations Ri establish a link between these elements.

Advantageously, the historical relations Rh and the historical links Lh are representative of the fusion strategy applied to the initial elements Ei via the function elements to obtain the merged element Ec.

Advantageously, the fusion history Hc determined for each merged element Ec includes the initial history Hi of each initial element Ei from which the merged element Ec originates. Such an inclusion allows a non-loss of information during multiple mergers. An example of history Hc is illustrated by FIGS. 6 and 7 for a merged element Ec obtained following two merging of elements. The first merge is shown in Figure 7 and the second merge in Figure 6.

The first merge made it possible to obtain the merged entity Ec "Man: John Dow", from the initial entities "Man: John Dow" and "Person: J. Dow". Box C1 of the Ec history gives information relating to the first merger.

In particular, the part of the merge history Hc relating to the first merge comprises: the merged element Ec "Man: John Dow" resulting from the first merge, the initial histories Hi 1 and Hi2 of the initial entities Chi from which the merged element Ec,

- the historical entities Ch and historical relations Rh relating to the merger strategy including: o a historical entity Ch of the "Merger function" type and value "keep full name", o a historical entity Ch of the "Merger strategy" type and value "*", o a historical entity Ch of type "Threshold" and value "0.8" o a historical entity Ch of type "Similarity measure" and value "detection of initials", o a historical relationship Rh of type "Relation" and value "A resource", o a historical relation Rh of type "Relation" and value "Merge function", o a historical relation Rh of type "Relation" and value "Compatibility function ”, and o a historical relation Rh of type “Relation” and value “Produced by”.

The second merge made it possible to obtain the merged entity Ec "Man: John Dow", from the initial entity "Person: J. Dow" and the entity resulting from the first merge "Man: John Dow" . Box C2 of the Ec history gives information relating to the second merger.

The part of the merge history Hc relating to the second merge includes in particular: the merged element Ec "Man: John Dow" resulting from the second merge, the histories of the entities from which the merged element originated Ec: initial history Hi3 and the part of the history relating to the first merger, and - the historical entities Ch and historical relations Rh relating to the merger strategy which are, in this case, the same as for the first merger.

All the types of historical entities Ch corresponding to the example of figures 6 and 7 are summarized in figure 8. The types of entities are in this example:

“Person” and “Man”, these entities referring to the initial and merged entities, the “Man” type being a subtype of the “Person” type.

"Merge strategy", "Similarity measure", "Threshold", "Merge function", these entities relating to the merging strategy, and

"Source", that entity relating to data sources.

Advantageously, each historical entity Ch is associated with at least one additional piece of data chosen from: a timestamp indicating the date and/or time of generation of the historical entity Ch (illustrated in the entity boxes of the examples of FIGS. 6 and 7) , data relating to the place of generation of the historical entity Ch, and data relating to the computer by which the historical entity Ch was generated, data relating to a user of the computer by which the historical entity Ch.

Optionally, the tracing method comprises a step 130 of acquiring a command to display the fusion history Hc obtained for a merged element Ec. For example, the command emanates from an operator and is acquired via the man-machine interface 22 of the computer 10. The command is, for example, issued with a view to verifying the origin of the merged elements Ec.

The plotting method comprises a step 140 of displaying the history of fusion Hc of said merged element Ec on a display, such as the display 24 of the computer 10. Depending on the history displayed, the final data Df are by example considered reliable or unreliable, including whether the original data source is considered reliable or unreliable.

Optionally, when the initial data Di relates to at least one piece of equipment, the method comprises a step 150 of implementing an action on the piece of equipment as a function of the final data Df obtained, and of at least one Hc fusion history. For example, if the initial data Di and final data Df relate to the maintenance of equipment, the maintenance action corresponding to the final data Df is carried out on the equipment, in particular if it has been determined that the final data Df are reliable in view of the fusion histories.

Thus, the present method makes it possible, via the determination of a fusion history Hc, to improve the understanding and reliability of the data resulting from a data fusion, as well as understanding itself as an operation on these data. This facilitates, thus, the exploitation of the merged data.

The present method thus applies particularly to the technical field of information processing, with application for example to decision support or to the analysis of situations or complex systems, in particular when the information to be merged comes from various heterogeneous sources of information. In particular, many military and civilian applications use analysis and decision systems based on information from a plurality of heterogeneous sources, such as for example a plurality of sensors (e.g. perimeter surveillance, detection and classification of targets , ammunition self-guiding for the military field, for the civilian field: industrial maintenance, weather forecasting, remote sensing).

One application is, for example, the analysis of data from heterogeneous sources for the collection of evidence. In this context, semantic fusion is used to merge heterogeneous data from different data sources. The traceability allowed by this process makes it possible to make the evidence collected more reliable.

Another application is, for example, the study of information exchanged on social networks where information of different quality levels is used. In particular, different levels of credibility are attributed to different actors in social networks. Thus, when using information relayed on social networks, this process makes it possible to trace the origin of the information, which makes it possible to assess its reliability.

The present method makes it possible in particular to (i) trace the various mergers carried out (ii), follow the steps carried out while keeping the order of their realizations, (iii) have proof of the steps carried out, and (iv) understand the steps carried out with a need for interpretability.

In one mode of implementation, the method makes it possible to keep a history of all the mergers carried out. The traces of the mergers are, for example, represented using conceptual graphs with a proper ontology which is composed of (i) sub-type and super-type of the entity which has been merged; (ii) an ontology specific to the traceability representing information related to the fusion, and (iii) an ontology specific to the sources of information. The entities and relations of the history graph, representing the history of the different merges for a given entity, make it possible to represent the similarity function and the associated threshold, the merge function and the compatibility function associated with the merges. These traces are represented for each entity or relationship as an additional property that represents its history. Each piece of information is preferably timestamped at the time it is created to keep a safe and unambiguous record. The present method proposes to follow the process of semantic fusion of heterogeneous data by adding to each merged entity a history in the form of a conceptual graph with its own formalism. This solution makes it possible to follow the evolution of the fusion of each concept and relations. Adding this history to each entity facilitates the history and readability of the information by knowing for each entity where it comes from. The fact of using the same formalism (conceptual graphs) and of storing the history information in the same form as the history itself allows, on the one hand, a homogeneity of the information and, on the other hand, on the other hand, if it is desired to condense the history (history graph) or to merge two histories (two or more history graphs), to define similarity and fusion functions specific to the history graph.

Thus, the present method has in particular the following advantages: (i) the trace of the various fusions carried out; (ii) the trace of all fusion parameters and strategy; (iii) the use of conceptual graphs to represent the trace (same way of representing the knowledge as the information itself which will be merged) makes it possible to merge traces if they are redundant to render traceability information itself condensed; (iv) the process is generic and adaptable to any use case; (v) the process allows an operator to understand the different steps that led to the synthetic information.

Those skilled in the art will understand that the embodiments described above can be combined to form new embodiments provided that they are technically compatible. In particular, variants of representation of the initial, final and historical data are compatible with the present method. In addition, the examples described relate to the merging of entities, but the present method is also compatible with the merging of relations. In addition, the types and values of the entities and relationships mentioned are not limiting and must be adapted according to the intended application.

Claims

1. A method of digitally plotting data from multiple data sources, the method being computer-implemented and comprising the steps of: a. obtaining initial data (Di) from several data sources, each initial data (Di) comprising initial elements (Ei) and initial links (Li), each initial element (Ei) having a type and a value, each element initial (Ei) belonging to a first category of elements, called initial entities (Ci), or to a second category of elements, called initial relations (Ri), the initial links (Li) connecting initial entities (Ci) between they via an initial relation (Ri), b. determination of a final data (Df) as a function of at least two initial data (Di) following the fusion of initial elements (Ei) of said initial data (Di) with each other according to a fusion strategy, the initial elements ( Ei) merged belonging to the same category of elements and making it possible to obtain a merged element (Ec), the merged element (Ec) having a type and a value, and c. determination of a fusion history (Hc) for each merged element (Ec), the fusion history (Hc) comprising the type and the value of each initial element (Ei) from which the merged element (Ec) originated, the type and value of the merged element (Ec) and data relating to the merging strategy.

2. Method according to claim 1, in which the merging strategy implements functions (F), each function (F) being represented by a set of function elements each comprising a type and a value, the history of fusion (Hc) comprising historical elements (Eh) and historical links (Lh), each historical element (Eh) having a type and a value, the historical elements (Eh) belonging to: a. a first category of historical elements called historical entities (Ch), grouping together the initial elements (Ei), the merged element (Ec) and the function elements representative of the functions (F), and b. a second category of historical elements, called historical relations (Rh), historical links (Lh) linking historical entities (Ch) to each other via historical relations (Rh).

3. Method according to claim 2, in which the historical relations (Rh) and the historical links (Lh) are representative of the merging strategy applied to the initial elements (Ei) via the function elements to obtain the merged element (Ec ).

4. Method according to any one of claims 1 to 3, in which each initial element (Ei) also has an initial history (Hi) comprising the type and the value of the initial element (Ei), as well as identification data of the data source from which the initial element (Ei) originated, the fusion history (Hc) determined for each merged element (Ec) including the initial history (Hi) of each initial element ( Ei) from which the merged element (Ec) originates.

5. Method according to any one of claims 1 to 4, in which each historical entity (Ch) is associated with at least one complementary datum chosen from: a. a timestamp indicating the date and/or time of generation of the historical entity (Ch), b. data relating to the place of generation of the historical entity (Ch), c. data relating to the computer by which the historical entity was generated (Ch), and d. data relating to a user of the computer by which the historical entity (Ch) was generated.

6. Method according to any one of claims 1 to 5, in which the initial data (Di) of at least one data source come from measurements made by a sensor.

7. A method according to any of claims 1 to 6, wherein the method comprises: a. a step of acquiring a display command for the merging history (Hc) determined for a merged element (Ec), and b. a step of displaying the fusion history (Hc) of said merged element (Ec) on a display. 17

8. Method according to any one of claims 1 to 7, in which the initial data (Di) relate to at least one piece of equipment, the method comprising a step of implementing an action on the piece of equipment as a function of the final data(s) (Df) obtained, and at least one melting history (Hc).

9. Computer program product comprising a readable information medium, on which is stored a computer program comprising program instructions, the computer program being loadable on a data processing unit and implementing a method according to any one of claims 1 to 8 when the computer program is implemented on the data processing unit.

10. Readable information medium on which is stored a computer program product according to claim 9.