A method for predicting a time extension period and related electronic device

Abstract

Disclosed is a method, performed by an electronic device, for predicting a time extension period. The method comprises obtaining a booking request for a shipment associated with a user. The method comprises obtaining, based on the booking request, historical data associated with the user. The method comprises predicting, based on the historical data and the booking request, a time extension period e.g., by applying a prediction model to the historical data. The method comprises providing, based on the predicted time extension period, a booking response for the user.

Description

[0001] A METHOD FOR PREDICTING A TIME EXTENSION PERIOD AND RELATED

[0002] ELECTRONIC DEVICE

[0003] The present disclosure pertains to the field of transport and freight. The present disclosure relates to a method for predicting a time extension period and related electronic device.

[0004] BACKGROUND

[0005] Shipping of items involves arrival and / or subsequent departure of items at terminal (e.g., port). Shipping of items via terminals (e.g., ports) involves managing Demurrage and / or Detention. It may be difficult for users to effectively manage Demurrage and / or Detention for a shipment.

[0006] SUMMARY

[0007] A user booking a shipment of items may select a time extension period for Demurrage and / or Detention (D&D) of items being shipped. However, it is difficult for a booking agent to determine and propose adequate time extension periods for the user to select a time extension period when booking the shipment.

[0008] Accordingly, there is a need for an electronic device and a method for predicting a time extension period, which mitigate, alleviate, or address the shortcomings existing and may allow for a more accurate, robust, and / or user-centric (e.g. personalized) prediction of the time extension period provided to a user.

[0009] Disclosed is a method, performed by an electronic device, for predicting a time extension period. The method comprises obtaining a booking request for a shipment associated with a user. The method comprises obtaining, based on the booking request, historical data associated with the user. The method comprises predicting, based on the historical data and the booking request, a time extension period e.g., by applying a prediction model to the historical data. The method comprises providing, based on the predicted time extension period, a booking response for the user.

[0010] Disclosed is an electronic device comprising memory circuitry, processor circuitry, and an interface, wherein the electronic device is configured to perform any of the methods disclosed herein. Disclosed is a computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by an electronic device cause the electronic device to perform any of the methods disclosed herein.

[0011] It is an advantage of the present disclosure that the disclosed electronic device and method provides a predicted time extension period for the user, by taking into account the historical data associated with the user. In other words, the disclosed predicted time extension period is personalized for the user and for the shipment, and thereby more optimal. This may lead to a more efficient handling of resources (such as equipment, transport, machinery in a port). Further, the disclosed method and electronic device benefit from a characterization of historical data patterns to provide a more accurate time extension in the booking response to the user. This may allow the user to minimize and / or prevent D&D penalties (e.g., fees).

[0012] BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The above and other features and advantages of the present disclosure will become readily apparent to those skilled in the art by the following detailed description of exemplary embodiments thereof with reference to the attached drawings, in which:

[0014] Fig. 1 is a diagram illustrating schematically an example process where the disclosed method for predicting a time extension period is carried out by an example electronic device according to this disclosure,

[0015] Figs. 2A-B are representations illustrative of example decision trees according to this disclosure,

[0016] Figs. 3A-B show a flow-chart illustrating an exemplary method, performed by an electronic device, for predicting a time extension period according to this disclosure, Fig. 4 is an example user interface representation of example outputs according to this disclosure, and

[0017] Fig. 5 is a block diagram illustrating an exemplary electronic device according to this disclosure.

[0018] DETAILED DESCRIPTION

[0019] Various exemplary embodiments and details are described hereinafter, with reference to the figures when relevant. It should be noted that the figures may or may not be drawn to scale and that elements of similar structures or functions are represented by like reference numerals throughout the figures. It should also be noted that the figures are only intended to facilitate the description of the embodiments. They are not intended as an exhaustive description of the disclosure or as a limitation on the scope of the disclosure.

[0020] In addition, an illustrated embodiment needs not have all the aspects or advantages shown. An aspect or an advantage described in conjunction with a particular embodiment is not necessarily limited to that embodiment and can be practiced in any other embodiments even if not so illustrated, or if not so explicitly described.

[0021] The figures are schematic and simplified for clarity, and they merely show details which aid understanding the disclosure, while other details have been left out. Throughout, the same reference numerals are used for identical or corresponding parts.

[0022] A shipment can be seen as shipping (e.g., transportation, such as transportation from a first location to a second location) of one or more items (e.g., goods). A shipment for example can be seen as shipment of a container, e.g., a container comprising items. A shipment may involve one or more modes of transportation, such as ocean carrier, land carrier, and / or air carrier.

[0023] Demurrage can be seen as the time from which a container arrives at a terminal and / or a port to the time the container is collected from the terminal and / or port to be transported to a recipient and / or destination (e.g., for unloading of at least a part of the container).

[0024] Detention can be seen as the time from which the container is collected from the terminal and / or port (e.g., for unloading) until the container is returned to the terminal and / or port (e.g., after unloading). For example, the returned container may comprise fewer items than prior to transportation from the terminal and / or port towards the recipient. In some examples, the returned container may be empty. Demurrage and / or detention may be referred to herein as D&D.

[0025] A booking request may be seen as request for booking of a shipment. For example, a booking request is provided by a user, possibly via a booking platform. In some examples, the user may be a booking agent. The booking request for example comprises information associated with the booking. In some examples, the booking request can be seen as a request for shipment of one or more items (e.g., goods). The booking request is, for example, based on user input provided by the user. In one or more examples, the booking request comprises a user code, country code, commodity code and contract code. The user code, country code, commodity code and / or contract code can be seen as attributes (e.g., features) of a shipment. The user code is for example indicative of or associated with a user. For example, the user code can be seen as a user identifier associated with a user, such as uniquely associated with the user. The user identifier may for example comprise a code (e.g., comprising one or more letters and / or numbers, e.g., alphanumeric code) indicative of a user identity. The country code is for example indicative of a country. For example, the country code comprises one or more letters and / or numbers, e.g., alphanumeric code indicative of a country. The commodity code is for example indicative of a commodity. For example, the commodity code may be seen as a commodity number (e.g., Harmonized System (HS) code). The commodity code for example comprises a code (e.g., alphanumeric code, a numerical code, a string, etc). The contract code is for example indicative of the contract associated with a shipment. In other words, the contract code can be indicative of the contract (e.g., the agreement) under which the shipment (e.g., the historical shipment) associated with the user was carried out. The contract code for example comprises a code (e.g., alphanumeric code, a numerical code, a string, etc). The contract code may for example be a contract number.

[0026] The user is for example a user of an electronic device, such as an electronic device communicatively coupled with the electronic device disclosed herein (e.g., electronic device 300 of Fig. 5).

[0027] The user may select a time extension period at the time of booking a shipment, such as when providing the booking request (e.g., to electronic device 300 of Fig. 5). For example, the user may select a time extension period to allow more time for D&D, including loading and / or unloading of the container. For example, the time extension period (e.g., free time extension) can be seen as a time period which may for example be selected by a user, e.g., for a given shipment. Time extension period can be seen as an extension of the D&D cut-off time, such as the last day of D&D detention before incurring a fee. In other words, the time extension period may be seen as an extension of the D&D time period (e.g., free time) such that a longer period of D&D may occur before a fee is incurred (such as to the user). Stated differently, the time extension period can for example be seen as D&D free time days.

[0028] Historical data associated with the user may be obtained based on the booking request. Historical data can for example be seen as previous (e.g., past) data associated with the user (e.g., one or more shipments of a user). In some examples, the historical data can be seen as user specific historical data. For example, the historical data can be seen as user specific historical D&D data. In some examples, the historical data associated with the user is processed using a predication model (such as a machine learning prediction model) to predict (e.g., derive) an optimal time extension period for the user and for the shipment.

[0029] Stated differently, the time extension period can be seen as an extension time slot, such as a D&D slot, such as a D&D time slot. In some examples, the predicted time extension period can be seen as a personalized predicted time extension period. The predicted time extension period can for example be provided (e.g., in a booking response) to the user (e.g., end user, booking agent, customer) while the user is booking a shipment. For example, the predicted time extension may be provided to a user of the electronic device (e.g., via interface 303 of Fig. 5, e.g. via a user interface illustrated in Fig. 4). The term “time extension period” may be used interchangeable with the term “time extension parameter” in this disclosure.

[0030] The user selecting a time extension period, such as based on the predicted time extension period, may help the user in mitigating the risk of unplanned delays (which may be costly) caused by e.g., Demurrage and Detention (D&D), applicable when stages of the shipment process exceed time periods agreed (e.g., agreed between a consignee and consignor prior to a shipment). However, the consignee may not be able mitigate the risk of unforeseen extension of times e.g. for D&D. Therefore, when providing a booking request, the user may select an inappropriate (e.g., too short) time extension period, resulting in one or more of occurrences: the carrier equipment (e.g., container) being held in the port for longer than agreed, and the carrier equipment not being returned to the port (e.g., after being collected) within the agreed time period. These possible events may respectively incur demurrage or detention fees (e.g., from the terminal authority) for the user. This may lead to disputes, revenue leakages, and eventually dissatisfaction of the user.

[0031] The present disclosure proposes to predict the time extension period so that a booking response can be provided based on the predicted time extension period determined for the user, enabling the user to select a more appropriate time extension period. In other words, the method aims to provide improved accuracy of predicted D&D time extension periods. This may reduce the risk of fees relating to D&D, and associated dissatisfaction of the user. Fig. 1 is a diagram illustrating schematically an example process where the disclosed method for predicting a time extension period is carried out by an example electronic device according to this disclosure, such as by the electronic device disclosed herein illustrated by electronic device 300 (e.g. electronic device 300 of Fig. 5). Fig. 1 can for example be seen as showing an end to end booking flow involving the electronic device 300 (e.g., a time extension period predictor).

[0032] Fig. 1 shows an example booking system 1 . The booking system 1 can be for example a self-service instant booking system (SSIB). Booking system 1 comprises a booking engine 6, the electronic device 300 disclosed herein (such as the electronic device 300 of Fig. 5), and a database 16.

[0033] Fig. 1 shows a user 2, such as a user 2 of the booking system 1 . As an example, the user 2 may be booking (e.g., requesting) a shipment via booking engine 6 (e.g., a booking portal). For example, booking the shipment comprises booking, such as requesting, a time extension period (e.g., D&D free time days). The user 2 may provide a booking request 4 to the booking engine 6. The booking engine 6 can provide a user interface through which the user 2 may input to and / or receive information from the electronic device 300. In some examples, the user 2 may provide a booking request 4 directly to the electronic device 300. The booking engine 6 may provide (e.g., based on the booking request 4) a request 8 for a predicted time extension parameter.

[0034] The booking request 4 for example comprises a user code, a country code, a commodity code and / or a contract code. For example, the booking request 4 comprises information indicative of the user 2, e.g. the user code which can be seen as a user identifier. For example, the booking request 4 comprises a user code (e.g., user identifier) indicative of the user 2.

[0035] The electronic device 300 comprises a processor circuitry 302 configured to perform the disclosed technique. The database 16 may for example be configured to store historical data associated with the user 2. The electronic device 300 can provide to database 16 a request 14 for historical data associated with the user 2. Upon receiving the request 14 for historical data associated with the user 2, the database 16 may provide historical data 18 associated with the user 2. In other words, the electronic device 300 may for example be configured to obtain (e.g., fetch, e.g., via a booking portal) historical data 18 associated with the user (e.g., from database 16) based on the booking request, such as based on one or more elements of the booking requests, such as the user code provided in the booking request.

[0036] The database 16 is for example an electronic data storage. In some examples, the database 16 is a remote data storage part of an external electronic device (e.g., external server), such a different electronic device to electronic device 300. In some examples, the electronic device 300 comprises the database 16. For example, memory circuitry (such as memory circuitry 301 of Fig. 5) of the electronic device 300 may comprise the database 16.

[0037] For example, the electronic device 300 is configured to train and / or run the prediction model (e.g., via the processor circuitry 302, such processor circuitry 302 of Fig. 5). In some examples, the electronic device 300 comprises a predictor engine 50 configured to execute the prediction model disclosed herein. The electronic device 300 and / or predictor engine 50 is for example configured to predict, based on the historical data 18 and the booking request 4, a time extension period (e.g., by applying the prediction model 50 to the historical data 18). In some examples, the predictor 50 comprises a training system 10 and / or a classifier 12. The predictor engine 50 is for example configured to train (e.g., based on historical data, such as historical data associated with user 2) the prediction model (such as a machine learning prediction model, such as a decision tree classifier). In other words, the electronic device 300 may be configured to classify using classifier 12 the historical data 18 for provision of a predicted time extension period 20 (e.g., by applying the prediction model, such as a decision tree classifier to the historical data 18 associated with the user 2). In other words, the electronic device 300 is for example configured to predict a time extension period for the user, (e.g. a free time extension period).

[0038] In some examples, the electronic device 300 is configured to classify the predicted output into one or more time slots. The one or more time slots may be seen as configurable time slots. In other words, the predicted time extension period may comprise one or more time slots for selection by the user. Stated differently, for example the one or more time slots provided in the booking response are configured (e.g., customised), such as for the user so that the user can select a time slot amongst the provided personalized time slots. For example, the one or more time slots comprise Slot 1 : Day 1 to Day 5, Slot 2: Day 6 to Day 10, and / or Slot 3: Day 11 to Day 15. For example, the time extension parameter may comprise one or more time slots (e.g., Slot 1 , Slot 2, Slot 3, etc.). In one or more examples, the electronic device 300 is configured to predict a time extension period, such as one or more time slots based on the historical data 18 associated with the user. In one or more examples, the electronic device 300 is configured to predict one or more time extension periods, such as one or more time slots based on the historical data 18 associated with the user. Stated differently, predicting the time extension period can be seen as predicting the slot information based on the historical data, e.g., historical records, associated with the user.

[0039] The electronic device 300 may be configured to provide (such as via interface 303 shown in Fig. 5) a booking response 20 based on the predicted time extension period 20 to the booking engine 6 or to the user 2. The booking engine may provide a booking response 22 to the user 2. The booking response 20, 22 for example comprises the predicted time extension period 20. For example, the booking response 20, 22 can be seen as a time extension response. In other words, the predicted free-time period may for example be provided, e.g. transmitted and / or displayed, to the user via the booking engine (e.g. running a booking portal, e.g., SSIB). In some examples, the predicted time extension parameter can be seen as a time extension period adapted to the user based on the user historical data.

[0040] In some examples, the electronic device 300 is configured to predict the time extension period, such as D&D time extension days, based on the historical data 18. In some examples, the predicting the time extension period comprises analysing the historical data to train the prediction model and to run the prediction model on the historical data 18. For example, the booking response 20, 22 comprises a time slot representative of the predicted time extension period. The booking response 22 can be seen as enabling the user 2 to select a personalized time slot (e.g., a “best” time slot) for D&D. For example, the user 2 may select a time extension period based on the booking response.

[0041] Fig. 2A is a representation illustrating an example decision tree 60 according to this disclosure. The decision tree 60 is associated with a decision tree classifier. In one or more example methods, the method comprises applying a prediction model to historical data associated with a user to determine the time extension period(s) disclosed herein. In some examples, the prediction model is a decision tree classifier configured to take as input the historical data associated with the user and to output the predicted time extension period(s). The historical data for example comprises one or more of: one or more attributes, one or more labels associated with each attribute, a count for each attribute, and one or more time extension parameters. The one or more attributes (e.g., A and B) can for example be seen as a type of features of the shipment. For example, the one or more attributes may be country (e.g., destination country or departure country), contract code, commodity code, user code, etc. In some examples, the one or more attributes may be seen as decision variables (e.g., of a decision tree). The one or more attributes can for example be seen as decision parameters. For example, the one or more attributes may be associated with one or more labels.

[0042] The one or more labels (e.g., illustrated by A1 -A7 and B1 -B2 in decision tree 60) associated with each attribute (e.g., illustrated by A and B) can be seen as values of the one or more attributes of the shipment. For example, when the attribute is country, a label may be “India, United Kingdom, Denmark, United States, or Japan” etc. For example, when the attribute is a contract code, a label may be a contract code of a shipment (e.g., “57”), such as of a historical shipment associated with the user.

[0043] The count for each label associated with an attribute can be seen as the number of times a shipment, such as a historical shipment, associated with the user has been associated with a given label of an attribute. The count is for example associated with a label (e.g., A3). For example, the count for each attribute can be seen as the number of shipments (e.g., of the historical data associated with the user) associated with a given label of the one or more attributes.

[0044] The historical data comprises time extension parameters (e.g., 1 -3). The one or more time extension parameters can be seen as parameters indicative of a time extension period. For example, a time extension parameter may comprise a value (e.g., decimal, integer, fraction, percentage, etc.)

[0045] In some examples, the method comprises determining, e.g., based on initial historical data associated with the user, the one or more attributes, one or more labels associated with each attribute, a count for each attribute, and / or one or more time extension parameters, by applying the prediction model to the initial historical data associated with the user. The initial historical data may be seen as booking data that has not been classified, thereby does not include attribute(s), count(s), label(s), actual previous time extension periods, etc. Fig. 2A shows a decision tree 60. The decision tree 60 is for example trained and / or generated based on historical data associated with a user. The decision tree shown in Fig. 2A can be seen as a generic decision tree. In some examples, the decision tree 60 can be seen as illustrating a decision tree classifier. In some examples, the decision tree 60 can be seen as illustrating a tree-structure classifier. The decision tree 60 comprises nodes. The decision tree 60 comprises internal nodes and leaf nodes. Node A and Nodes B can be seen as internal nodes of the decision tree 60. For example, internal nodes of the decision tree 60 can be seen as a node with, such as associated with, one or more child nodes. For example, an internal node can be a root node, such as node A. The root node and internal nodes can be seen as representative of attributes provided in the historical data (e.g. country, contract code, etc.). Leaf nodes can be seen as nodes of the decision tree 60 without, such as not associated with, a child node (e.g., nodes 1 -3). A leaf node is for example representative of the outcome of the prediction model, e.g. a class providing the predicted time extension period. In some examples, a leaf node may indicate the predicted time extension period. For example, nodes 1 , 2, and 3 are indicative of time slots 1 , 2, and 3 respectively, e.g. wherein slot 1 : Day 1 to Day 5, slot 2: Day 6 to Day 10, and slot 3: Day 11 to Day 15.

[0046] The decision tree classifier for example comprises one or more branches (e.g. one or more edges). Each branch can be seen as connecting two nodes of the decision tree 60. A branch is for example representative of a decision rule. A decision rule can be seen as a set of conditions that classify the historical data, e.g., historical records. In other words, each branch represents the outcome of the decision rule between the parent node and the child node, and each leaf node represents a class label (e.g. decision taken after computing all attributes forming the nodes of the decision tree). In other words, the paths from root to leaf may represent classification rules. For example, the decision tree (e.g., a classification tree) is a tree in which each internal (non-leaf) node is labelled with an attribute (e.g. input feature). For example, the branches coming from a node labelled with an attribute are labelled with each of the possible values of the target attribute or the branches leads to a subordinate decision node on a different attribute. For example, each leaf of the tree is labelled with a class or a probability distribution over the classes, signifying that the data set has been classified by the decision tree into either a specific class, or into a particular probability distribution (which may be skewed towards certain subsets of classes). Each class is for example represented by a label (e.g. a time extension label), such as 1 , 2, 3 in Fig. 2A. As can be seen in Fig. 1 , 7 nodes (node A and nodes B) of the decision tree 60 are representative of attributes (e.g., features) A and B of the one or more attributes (e.g., features). A and B are for example attributes (e.g., different attributes) of a shipment associated with the user. The node representative of the attribute A may be referred to as node A. The node representative of the attribute B may be referred to as node B. Child nodes of node A are connected to node A by branches A1 -A7 representative of corresponding labels for attribute A. In other words, attribute A is given as value or label any of A1 -A7 in the historical data obtained. Nodes B can be seen as child nodes of node A (the root node of decision tree 60.

[0047] The decision tree 60 comprises branches representative of A1-A7 and B1-B2. For example, labels A1-A7 are associated with attribute A. The labels A1 -A7 can be seen as unique labels (e.g., each label may be different). In other words, branches representative of attributes A1-A7 can be seen as branching off from node A. The branches representative of attributes A1 -A7 can be seen as first branches of the decision tree 60.

[0048] The labels B1-B2 can be seen as branches of nodes B. In other words, branches representative of attributes B1 -B2 can be seen as branching off from nodes B. The branches representative of attributes B1 can be seen as first primary branches of the decision tree 60. The branches representative of attributes B2 can be seen as first secondary branches of the decision tree 60.

[0049] Fig. 2B is a representation illustrative of an example decision tree 80 according to this disclosure. The decision tree 80 has a similar structure as the decision tree 60 shown in Fig. 2A. The decision tree 80 is associated with a decision tree classifier. In other words, the decision tree 80 comprises internal nodes, branches, and leaf nodes. The decision tree 80 is for example generated based on the historical data associated with a user shown in Table 1 below.

[0050] Table 1

[0051] Table 1 shows historical data associated with a user. For example, each row of Table 1 is associated with a historical shipment associated with the user. Table 1 comprises attributes and time extension parameters. The attributes shown in Table 1 are the contract codes (e.g., 47 and 57) and countries (e.g., Denmark, India, etc.). As shown in Table 1 , each contract code can be associated with one or more shipments. Time extension parameters shown in Table 1 are values between 1 -3. For example, the time extension parameters shown in Table 1 can be seen as time slots, e.g., time extension parameters.

[0052] The root node (e.g., node A of Fig. 2A) of decision tree 80 is representative of the country (e.g., destination country) of a shipment associated with a user. The country is for example an attribute of a shipment associated with the user. The node A shown in Fig. 2A for example corresponds with the root node representative of the country attribute of Fig.

[0053] 2B.

[0054] Decision tree 80 comprises branches (e.g., branches A1 -A7 of Fig. 2A) linking the root node (e.g., node A of Fig. 2A) representative of the country attribute and the internal nodes (e.g., nodes B of Fig. 2B) representative of contract code attribute. The branches linking the root node and the internal nodes can be seen as first branches. As shown in Fig. 2B, the first branches can be representative of countries (e.g., India, United Kingdom, Denmark, United States, Japan, Indonesia, China, etc.) corresponding to the historical data provided in Table 1 . It is to be noted that for the country code UK, the decision tree classifier provides directly option 2 as the perfectly classified option. For example, the United Kingdom shown in Table 1 is only associated with 1 time extension parameter (slot 2). This is shown in decision tree 80 as the branch (e.g., first branch A7) directly links from the root node representative of the attribute country, to the leaf node representative of a time extension period (slot 2).

[0055] The decision tree 80 comprises internal nodes (e.g., internal nodes B of Fig. 2A) representative of the contract code of a shipment associated with a user. The contract code may be seen as the contract code (e.g., contract number) of the contract (e.g., the agreement) under which the shipment (e.g., the historical shipment) associated with the user was carried out.

[0056] The decision tree 80 comprises branches (e.g., branches B1-B2) linking the internal nodes (e.g., nodes B of Fig. 2A) representative of the contract code and the leaf nodes representative of the time extension parameters (e.g. slot 1 , 2, or 3). The branches linking internal nodes and the leaf nodes can be seen as second branches. In some examples, each internal node may be associated with one or more second branches linking the internal node to a child node (such as a leaf node, e.g., the node representative of the time extension parameter). As shown in Table 1 and Fig 2B there are two labels (47 and 57) associated with the contract code. In other words, the value for a contract code can be 47 or 57. The internal nodes of decision tree 80 comprise 2 second branches, each linking the internal node to a child node. The second branches associated with each internal node of decision tree 80 are each representative of a given contract code of the historical data, e.g., 47 or 57. The decision tree 80 comprises one or more leaf nodes representative of one or more corresponding time extension parameters. For example, the decision tree 80 comprises leaf nodes representative of time parameter extensions: slot 1 , 2, or 3. The time extension parameter of the leaf node can for example be seen as a classified time extension parameter.

[0057] For example, predicting, based on the historical data (e.g., Table 1) and the booking request, the time extension period comprises applying a prediction model (e.g., decision tree classifier, such as decision tree 80) to the historical data.

[0058] In some examples, predicting, based on the historical data (e.g., Table 1 ) and the booking request, the time extension period comprises determining whether the historical data provides the same time extension parameter for all data elements of the historical data. In other words, for example, the electronic device determines if all rows of the historical data, e.g. in Table 1 , have the same time extension parameter. For example, upon determining that all rows of the historical data (e.g., rows of a Table, e.g., of Table 1) have the same time extension parameter, the electronic device makes the current node the leaf node. Stated differently, for example, all rows belong to the same class when it is determined that all rows of the historical data, e.g. in Table 1 , have the same time extension parameter. In other words, upon determining that the time extension parameters for all data elements of the historical data is the same, the historical data can be seen as classified, e.g. perfectly classified. For example, upon determining the time extension parameters for all data elements of the historical data is the same (e.g., when the historical data has been classified), the time extension parameter associated with the attributes is selected as the predicted time extension period. In some examples, the method comprises, upon determining that the time extension parameters for all data elements of the historical data is the same, generating a leaf node representative of the time extension parameter.

[0059] In some examples, the method comprises, upon determining that all data elements of the historical data (e.g., rows of a Table, e.g. of Table 1) do not have the same time extension parameter, checking whether at least one attribute is provided in the historical data (e.g., Table 1 ). For example, when all data elements shown in Table 1 (country and contract code) are not associated with the same time extension parameter, the electronic device determines whether the historical data provides at least one attribute. For example, the electronic device verifies a count of parameters from the historical data (such as columns of a Table, e.g. Table 1 , for example retrieved from a database).

[0060] In some examples, upon determining that no attribute is included in the historical data, the electronic device makes the current node (e.g., root node or internal node) a leaf node e.g., with the leaf node being representative of the most frequently occurring time extension parameter of the historical data. In other words, the current node is given the class having the highest occurrence as its label. For example, when the historical data comprises time extension parameters and does not comprise attributes (e.g., country and / or contract code), the most frequently occurring time extension parameter may be taken (e.g., selected) as the predicted time extension parameter.

[0061] In some examples, when at least one attribute is provided in the historical data, the method comprises calculating an information gain of the at least one attribute. In some examples, the historical data includes one or more attributes, and the information gain is calculated for each attribute.

[0062] In some examples, the method comprises selecting a first attribute providing the highest information gain amongst the one or more attributes. In other words, for example, the historical data is divided into subsets based on the attribute for which the information gain is the highest (e.g., highest amongst the information gains of the attributes of the historical data).

[0063] Information Gain can be seen as indicative of how well a given attribute (e.g., feature) classifies the target classes. The attribute with the highest Information Gain can be seen as the most performant attribute (e.g., for classifying the historical data).

[0064] For example, the information gain may of an attribute of a given data set be expressed as:

[0065] Gain(S, A) = Entropy(S) - £ [p(S|A) . Entropy(S|A)] (1)

[0066] Where p(l) denotes a probability of total occurrence of a distinct value over total occurrence of all values, S denotes a set (e.g. entire data set of the historical data), A denotes an attribute, and n denotes a number of unique output values. The entropy of a dataset is the measure of disorder and / or impurities in the target attribute (e.g., feature) of the data set.

[0067] Entropy(S) = £ - p(l).lognp(l) (2) where I denotes a random variable representative of the occurrence of a unique label with respect to the output time extension period (e.g. slot) of the target attribute.

[0068] In some examples, the method comprises generating a decision tree (e.g., decision tree 80) based on the attribute for which the information gain is highest. In other words, the method can comprise generating a decision tree where the root node is representative of the attribute (referred to as first attribute) with the highest calculated information gain amongst the information gains calculated for the attributes of the historical data. In some examples, the method comprises generating a node (e.g., a child node of an internal node, e.g., of the root node) representative of the attribute (referred to as first attribute) with the highest information gain.

[0069] In some examples, the method comprises generating one or more nodes of a decision tree. For example, generating one or more (e.g., each) nodes of a decision tree may comprise determining which attribute has the highest information gain. In one or more examples, predicting the time extension period comprises, upon determining that the historical data comprises at least one attribute, selecting, as a root node of the decision, a first attribute providing the highest information gain amongst the one or more attributes. For example, the first attribute of the root node of decision tree 80 is country. For the decision tree 80, country can be seen as the attribute with the highest information gain (e.g., the first attribute). For example, the attribute of the child node of the root node is contract code. Contract code (in decision tree 80) has the second highest information gain. In other words, the attribute with the second highest information gain can be seen as the second attribute (e.g., the child node of the root node). For example, the internal node representative of contract code, shown in Fig. 2B, (e.g., the first internal non-root node, e.g., the child node of the root node) is representative of the second attribute (e.g., the attribute with the second highest information gain, e.g., contract code as seen in Table 1). The child node of the root node can be seen as a first internal non-root node. The child node of the first internal non-root node can be seen as a second internal non-root node.

[0070] In other words, the attribute of the root node may for example be the attribute with the highest calculated information gain (e.g., of the historical data). The attribute of the first internal non-root node may for example be the attribute with the second highest calculated information gain of the one or more attributes (e.g., of the historical data). In some examples, when the historical data comprises a third attribute with a calculated information gain lower than that of the second attribute, the electronic device generates a child node of the internal non-root node representative of the third attribute.

[0071] The process can be repeated until all attributes of the historical data are exhausted, and all data elements of the historical data are classified, e.g. to belong to the same class.

[0072] Figs. 3A-B show a flow diagram of an exemplary method 100, performed by an electronic device, for predicting a time extension period according to the disclosure. The method 100 is performed by an electronic device, such as the electronic device disclosed herein, such as electronic device 300 of Fig. 5.

[0073] The method 100 comprises obtaining S102 a booking request for a shipment associated with a user. The method 100 for example comprises receiving and / or retrieving a booking request for a shipment associated with a user, e.g. from a booking platform, and / or a user device. The booking request can for example be seen as a user specific request for a shipment associated with a user. The booking request for example can be seen as a request for a time extension period.

[0074] The method 100 comprises obtaining S106, based on the booking request, historical data associated with the user. In some examples, obtaining S106, based on the booking request, historical data associated with the user comprises retrieving and / or receiving the historical data, e.g. from a local storage and / or a remote storage. In some examples, obtaining S106, based on the booking request, historical data associated with the user comprises obtaining historical data from remote storage (e.g., from a database, such as database 16 of Fig. 1 ). In one or more example methods, obtaining S106 the historical data comprises obtaining S106A one or more of: one or more attributes, one or more labels associated with each attribute, a count for each attribute, and one or more time extension parameters indicative of one or more previous time extension periods associated with corresponding one or more attributes. In some examples, obtaining S106A one or more of: one or more attributes, one or more labels associated with each attribute, a count for each attribute, and one or more time extension parameters indicative of one or more previous time extension periods associated with corresponding one or more attributes comprises determining (e.g. generating) one or more of: the one or more attributes, one or more labels associated with each attribute, and a count for each attribute and one or more time extension parameters indicative of one or more previous time extension periods associated with corresponding one or more attributes. It may be envisaged that at an initial stage, initial historical data associated with the user includes booking data (e.g. historical booking data). In such a case, for example, the historical data is generated by electronic device by indexing the initial historical data to determine the one or more attributes, one or more labels associated with each attribute, a count for each attribute, and / or one or more time extension parameters. For example, the historical data is generated by indexing the initial historical data associated with the user.

[0075] The method 100 comprises predicting S108, based on the historical data and the booking request, a time extension period by applying a prediction model to the historical data. In some examples, the time extension period is predicted based on one or more of: one or more attributes, one or more labels associated with each attribute, a count for each attribute, one or more time extension parameters indicative of one or more previous time extension periods associated with corresponding one or more attributes, and the booking request. In other words, in some examples, the prediction model is applied to one or more of: one or more attributes, one or more labels associated with each attribute, a count for each attribute, one or more time extension parameters indicative of one or more previous time extension periods associated with corresponding one or more attributes, and the booking request.

[0076] The method 100 comprises providing S110, based on the predicted time extension period, a booking response for the user. In some examples, providing S110, based on the predicted time extension period, a booking response for the user may comprise providing the predicted time extension period for the user. For example, the booking response may comprise the predicted time extension period. The booking response can be seen as a response to the booking request for a shipment associated with a user. The booking response for example comprises the predicted time extension parameter.

[0077] In one or more example methods, the prediction model is a machine learning prediction model. In one or more example methods, the machine learning prediction model is a decision tree classifier. A decision tree classifier can be seen as machine learning classifier for classification of information (e.g., historical data associated with the user, e.g. initial historical data associated with the user). The decision tree classifier can have a tree-like structure where internal nodes may be represented by attributes (e.g., features). For example, each internal node represents a decision based on a given attribute. The leaf nodes of the tree-like structure of the decision classifier for example represent outcomes (e.g., outputs) of the decision tree classifier. In other words, the internal nodes correspond with decision rules based on given attributes and leaf nodes represent the resulting classifications. A decision tree classifier may be configured to generate a decision tree based on highest information gain. For example, during classification, a user input instance (e.g., commodity code, user code, contract code and / or country) may be guided down the tree following decision rules until it reaches a leaf node (e.g., time extension parameter). Figs. 2A-B illustrate a decision tree classifier according to this disclosure. The attribute may be seen as a decision variable and / or a decision parameter for the decision tree classifier.

[0078] In one or more example methods, predicting S108 the time period comprises training S108A, based on the historical data, the decision tree classifier. In one or more example methods, training S108A, based on the historical data, the decision tree classifier comprises classifying S108AA the historical data using the decision tree classifier (e.g. following any of S108B-I).

[0079] In one or more example methods, predicting S108 the time extension period comprises determining S108B whether the historical data provides the same time extension parameter for all data elements of the historical data. In other words, the historical data comprises one or more data elements (e.g. Table 1 ). In one or more example methods, predicting S108 the time extension period comprises upon determining that the historical data provides the same time extension parameter for all data elements of the historical data, taking S108C the value of the same time extension parameter as the predicted time extension period. For example, when the historical data provides the same time extension parameter (e.g. value, and / or period) for all data elements, then the time extension period provided in the time extension parameter of the historical is the one selected to be the predicted time extension period and provided for booking response.

[0080] In one or more example methods, predicting S108 the time extension period comprises upon determining that the historical data does not provide the same time extension parameter for all data elements of the historical data, determining S108D whether the historical data comprises at least one attribute. In some examples, the historical data comprises a plurality of attributes, such as a first attribute, a second attribute and optionally a third attribute, and optionally a fourth attribute etc. In one or more example methods, predicting S108 the time extension period comprises upon determining that the historical data does not comprise at least one attribute, taking S108E the time extension parameter having the highest occurrence as the predicted time extension period. For example the time extension parameter of the historical data having the highest occurrence is the one being the most frequent in the historical data. This is for example illustrated with Fig. 2B.

[0081] In one or more example methods, predicting S108 the time extension period comprises, upon determining that the historical data comprises at least one attribute (e.g. one or more attributes), determining S108F an information gain for each attribute (e.g. of the one or more attributes). This is for example illustrated with Fig. 2B.

[0082] In one or more example methods, predicting S108 the time extension period comprises, upon determining that the historical data comprises at least one attribute, selecting S108G a first attribute providing the highest information gain amongst the one or more attributes. The information gain is illustrated e.g. by Equation (1 ).

[0083] In one or more example methods, predicting S108 the time extension period comprises, upon determining that the historical data comprises at least one attribute, generating S108H, based on the first attribute, a first decision tree having the first attribute as a root node and one or more leaf nodes representative of one or more predicted time extension periods. An example first decision tree may be seen as decision tree of Figs. 2A-B with the root node, non-root internal nodes and leaf nodes. The first decision tree can be seen as the decision tree with the first attribute as a root node. In other words, the first decision tree can be seen as a decision tree where the root node is representative of the attribute with the highest information gain. For example, a second decision tree can be seen as a decision tree, such as a sub-tree of the first decision tree, thereby having the same root node as the first decision tree.

[0084] In one or more example methods, predicting S108 the time extension period comprises, upon determining that the historical data comprises at least one attribute, classifying S108I the historical data using the first decision tree. This may lead to additional internal non-root nodes of the first decision tree, as illustrated for Figs. 2A-B.

[0085] In one or more example methods, predicting S108 the time extension period comprises repeating one or more of steps S108B to S108I, until all the attributes amongst the one or more attributes of the historical data are processed. In some examples, classifying S108AA the historical data using the decision tree classifier comprises repeating one or more of steps S108B to S108I e.g., until all the attributes amongst the one or more attributes of the historical data are processed.

[0086] In one or more example methods, the method 100 comprises determining S104 whether the historical data is retrievable from storage. In one or more example methods, the method 100 comprises upon determining that the historical data associated with the user is retrievable from storage, retrieving S106B the historical data from the storage. In one or more example methods, the method 100 comprises upon determining that the historical data associated with the user is not retrievable from storage, refraining from retrieving the historical data (e.g. exiting method 100).

[0087] In one or more example methods, providing S110, based on the predicted time extension period, the booking response for the user comprises causing S110A a display of a user interface object representative of the predicted time extension period. In some examples, a user device may include a display that in response to receiving the booking response provided in S110, is configured to display the user interface object representative of the predicted time extension period

[0088] Fig. 4 is an example representation of example outputs according to this disclosure. For example, Fig. 4 shows a user interface that may be displayed on a user device including a display in response to receiving a booking response disclosed herein.

[0089] Fig. 4 shows a user interface object 92 representative of an example predicted time extension period provided in a booking response and an example time extension value table 90. For example, providing a booking response can be seen as causing the display of an electronic device (e.g., the interface 303 of electronic device 300 shown in Fig. 5) to display the user interface object 92 representative of the predicted time extension period. Stated differently, the user interface object 92 represents the predicted time extension period. The user interface object 92 is for example a drop-down menu providing a plurality of time extension periods predicted according to the disclosure. The user interface object 92 may for example enable a user to select a time extension period, e.g., based on the predicted time extension periods. In some examples, the predicted time extension period can be seen as a recommended time extension period. The user interface object 92 is representative of a predicted time extension period of +4 days (e.g., 4 day extension to the pre-existing time period). In other words, the electronic device may be configured to add 4 days (e.g., based on user input) to the 1 -12 free time period (e.g., 12 free days).

[0090] In the example of Fig. 4, the user interface object 92 is indicative of the value (e.g., price) incurred (e.g., by a port authority and / or the consignor) when a user selects a time extension period. The value incurred is for example displayed in a given currency (e.g., Indian Roubles (INR) as shown in Fig. 4).

[0091] The time extension value table 90 comprises information indicative of the value (e.g., price) of a time extension period. For example, the time period 1-12 can be seen as the D&D time period that does not incur additional cost. The time periods 13-15 and 15+ can be seen as time extension periods. The values in the table (0, 120,00, and 350,00) are values in Emirati Dirham (AED) currency (any currency may be applied).

[0092] Fig. 5 shows a block diagram of an exemplary electronic device 300 according to the disclosure. The electronic device 300 comprises memory circuitry 301 , processor circuitry 302, and an interface 303. The electronic device 300 is configured to perform any of the methods disclosed in Figs. 3A-B. In other words, the electronic device 300 is configured for predicting a time extension period.

[0093] In some examples, the electronic device 300 is a time extension prediction electronic device. In some examples, the electronic device 300 is a time extension prediction electronic system. In some examples, the electronic device 300 is a time extension predictor.

[0094] The electronic device 300 is configured to obtain (e.g., via memory circuitry 301 and / or interface 303) a booking request for a shipment associated with a user.

[0095] The electronic device 300 is configured to obtain (e.g., via memory circuitry 301 and / or interface 303), based on the booking request, historical data associated with the user.

[0096] The electronic device 300 is configured to predict (e.g., via processor circuitry 302), based on the historical data and the booking request, a time extension period by applying (e.g., via memory circuitry 301 and / or processor circuitry 302) a prediction model to the historical data. The electronic device 300 is configured to provide (e.g., via processor circuitry 302 and / or interface 303), based on the predicted time extension period, a booking response for the user.

[0097] The processor circuitry 302 is optionally configured to perform any of the operations disclosed in Figs. 3A-B (such as any one or more of: S102, S104, S106, S106A, S106B, S108, S108A, S108B, S108C, S108CA, S108D, S108E, S108F, S108G, S108H, S108I, S110, S110A). The operations of the electronic device 300 may be embodied in the form of executable logic routines (e.g., lines of code, software programs, etc.) that are stored on a non-transitory computer readable medium (e.g., the memory circuitry 301 ) and are executed by the processor circuitry 302).

[0098] Furthermore, the operations of the electronic device 300 may be considered a method that the electronic device 300 is configured to carry out. Also, while the described functions and operations may be implemented in software, such functionality may as well be carried out via dedicated hardware or firmware, or some combination of hardware, firmware and / or software.

[0099] The memory circuitry 301 may be one or more of a buffer, a flash memory, a hard drive, a removable media, a volatile memory, a non-volatile memory, a random access memory (RAM), or other suitable device. In a typical arrangement, the memory circuitry 301 may include a non-volatile memory for long term data storage and a volatile memory that functions as system memory for the processor circuitry 302. The memory circuitry 301 may exchange data with the processor circuitry 302 over a data bus. Control lines and an address bus between the memory circuitry 301 and the processor circuitry 302 also may be present (not shown in Fig. 5). The memory circuitry 301 is considered a non-transitory computer readable medium.

[0100] The memory circuitry 301 may be configured to store a booking request, historical data, time extension period, prediction model, booking response, one or more attributes, one or more labels associated with each attribute, a count for each attribute, one or more time extension parameters, information gain and / or user interface object in a part of the memory.

[0101] Embodiments of methods and products (electronic device) according to the disclosure are set out in the following items: Item 1 . A method, performed by an electronic device, the method comprising: obtaining a booking request for a shipment associated with a user; obtaining, based on the booking request, historical data associated with the user; predicting, based on the historical data and the booking request, a time extension period by applying a prediction model to the historical data; and providing, based on the predicted time extension period, a booking response for the user.

[0102] Item 2. The method according to item 1 , wherein the prediction model is a machine learning prediction model.

[0103] Item 3. The method according to any of the previous items, wherein the machine learning prediction model is a decision tree classifier.

[0104] Item 4. The method according to any of the previous items, wherein obtaining the historical data comprises obtaining one or more of: one or more attributes, one or more labels associated with each attribute, a count for each label associated with an attribute, and one or more time extension parameters indicative of one or more previous time extension periods associated with corresponding one or more attributes.

[0105] Item 5. The method according to any of items 3-4, wherein predicting the time extension period comprises training, based on the historical data, the decision tree classifier.

[0106] Item 6. The method according to item 5, wherein training, based on the historical data, the decision tree classifier comprises classifying the historical data using the decision tree classifier.

[0107] Item 7. The method according to any of the previous items, wherein predicting the time extension period comprises:

[0108] - determining whether the historical data provides the same time extension parameter for all data elements of the historical data; and

[0109] - upon determining that the historical data provides the same time extension parameter for all data elements of the historical data, taking the value of the same time extension parameter as the predicted time extension period. Item 8. The method according to any of the previous items, wherein predicting the time extension period comprises:

[0110] - upon determining that the historical data does not provide the same time extension parameter for all data elements of the historical data, determining whether the historical data comprises at least one attribute.

[0111] Item 9. The method according to any of the previous items, wherein predicting the time extension period comprises:

[0112] - upon determining that the historical data does not comprise at least one attribute, taking the time extension parameter having the highest occurrence as the predicted time extension period.

[0113] Item 10. The method according to any of the previous items, wherein predicting the time extension period comprises, upon determining that the historical data comprises at least one attribute: determining an information gain for each attribute; selecting a first attribute providing the highest information gain amongst the one or more attributes; and generating, based on the first attribute, a first decision tree having the first attribute as a root node and one or more leaf nodes representative of one or more predicted time extension periods; classifying the historical data using the first decision tree.

[0114] Item 11. The method according to any of the previous items, wherein predicting the time extension period comprises repeating one or more of steps to, until all the attributes amongst the one or more attributes of the historical data are processed.

[0115] Item 12. The method according to any of the previous items, the method comprising determining whether the historical data is retrievable from storage. Item 13. The method according to any of the previous items, wherein providing, based on the predicted time extension period, the booking response for the user comprises causing a display of a user interface object representative of the predicted time extension period.

[0116] Item 14. An electronic device comprising memory circuitry, processor circuitry, and an interface, wherein the electronic device is configured to perform any of the methods according to any of items 1 -13.

[0117] Item 15. A computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by an electronic device cause the electronic device to perform any of the methods of items 1 -13.

[0118] The use of the terms “first”, “second”, “third” and “fourth”, “primary”, “secondary”, “tertiary” etc. does not imply any particular order, but are included to identify individual elements. Moreover, the use of the terms “first”, “second”, “third” and “fourth”, “primary”, “secondary”, “tertiary” etc. does not denote any order or importance, but rather the terms “first”, “second”, “third” and “fourth”, “primary”, “secondary”, “tertiary” etc. are used to distinguish one element from another. Note that the words “first”, “second”, “third” and “fourth”, “primary”, “secondary”, “tertiary” etc. are used here and elsewhere for labelling purposes only and are not intended to denote any specific spatial or temporal ordering. Furthermore, the labelling of a first element does not imply the presence of a second element and vice versa.

[0119] It may be appreciated that Figs. 1 -5 comprises some circuitries or operations which are illustrated with a solid line and some circuitries or operations which are illustrated with a dashed line. The circuitries or operations which are comprised in a solid line are circuitries or operations which are comprised in the broadest example embodiment. The circuitries or operations which are comprised in a dashed line are example embodiments which may be comprised in, or a part of, or are further circuitries or operations which may be taken in addition to the circuitries or operations of the solid line example embodiments. It should be appreciated that these operations need not be performed in order presented. Furthermore, it should be appreciated that not all of the operations need to be performed. The exemplary operations may be performed in any order and in any combination. It is to be noted that the word "comprising" does not necessarily exclude the presence of other elements or steps than those listed.

[0120] It is to be noted that the words "a" or "an" preceding an element do not exclude the presence of a plurality of such elements.

[0121] It should further be noted that any reference signs do not limit the scope of the claims, that the exemplary embodiments may be implemented at least in part by means of both hardware and software, and that several "means", "units" or "devices" may be represented by the same item of hardware.

[0122] The various exemplary methods, devices, nodes and systems described herein are described in the general context of method steps or processes, which may be implemented in one aspect by a computer program product, embodied in a computer- readable medium, including computer-executable instructions, such as program code, executed by computers in networked environments. A computer-readable medium may include removable and non-removable storage devices including, but not limited to, Read Only Memory (ROM), Random Access Memory (RAM), compact discs (CDs), digital versatile discs (DVD), etc. Generally, program circuitries may include routines, programs, objects, components, data structures, etc. that perform specified tasks or implement specific abstract data types. Computer-executable instructions, associated data structures, and program circuitries represent examples of program code for executing steps of the methods disclosed herein. The particular sequence of such executable instructions or associated data structures represents examples of corresponding acts for implementing the functions described in such steps or processes.

[0123] Although features have been shown and described, it will be understood that they are not intended to limit the claimed disclosure, and it will be made obvious to those skilled in the art that various changes and modifications may be made without departing from the scope of the claimed disclosure. The specification and drawings are, accordingly, to be regarded in an illustrative rather than restrictive sense. The claimed disclosure is intended to cover all alternatives, modifications, and equivalents.

Claims

CLAIMS1 . A method, performed by an electronic device, the method comprising: obtaining a booking request for a shipment associated with a user; obtaining, based on the booking request, historical data associated with the user; predicting, based on the historical data and the booking request, a time extension period by applying a prediction model to the historical data; and providing, based on the predicted time extension period, a booking response for the user.

2. The method according to claim 1 , wherein the prediction model is a machine learning prediction model.

3. The method according to any of the previous claims, wherein the machine learning prediction model is a decision tree classifier.

4. The method according to any of the previous claims, wherein obtaining the historical data comprises obtaining one or more of: one or more attributes, one or more labels associated with each attribute, a count for each label associated with an attribute, and one or more time extension parameters indicative of one or more previous time extension periods associated with corresponding one or more attributes.

5. The method according to any of claims 3-4, wherein predicting the time extension period comprises training, based on the historical data, the decision tree classifier.

6. The method according to claim 5, wherein training, based on the historical data, the decision tree classifier comprises classifying the historical data using the decision tree classifier.

7. The method according to any of the previous claims, wherein predicting the time extension period comprises:- determining whether the historical data provides the same time extension parameter for all data elements of the historical data; andupon determining that the historical data provides the same time extension parameter for all data elements of the historical data, taking the value of the same time extension parameter as the predicted time extension period.

8. The method according to any of the previous claims, wherein predicting the time extension period comprises:- upon determining that the historical data does not provide the same time extension parameter for all data elements of the historical data, determining whether the historical data comprises at least one attribute.

9. The method according to any of the previous claims, wherein predicting the time extension period comprises:- upon determining that the historical data does not comprise at least one attribute, taking the time extension parameter having the highest occurrence as the predicted time extension period.

10. The method according to any of the previous claims, wherein predicting the time extension period comprises, upon determining that the historical data comprises at least one attribute: determining an information gain for each attribute; selecting a first attribute providing the highest information gain amongst the one or more attributes; and generating, based on the first attribute, a first decision tree having the first attribute as a root node and one or more leaf nodes representative of one or more predicted time extension periods; classifying the historical data using the first decision tree.

Images