Systems and methods for determining estimated time of arrival

1. A system for determining an estimated time of arrival (ETA), comprising:

at least one storage medium including a set of instructions; and

at least one processor configured to communicate with the at least one storage medium, wherein when executing the set of instructions, the at least one processor is directed to:

receive a start location and a destination from a user device;

obtain a global ETA model from a storage medium, wherein to generate the global ETA model, the at least one processor is further directed to:

obtain feature data related to an on-demand service;

determine a plurality of sub-ETA models;

obtain a parallel computing framework including a plurality of worker nodes, wherein each of the plurality of worker nodes is associated with a sub-ETA model of the plurality of sub-ETA models;

allocate the feature data to the plurality of worker nodes;

train the plurality of sub-ETA models based on the feature data; and

generate the global ETA model based on the plurality of trained sub-ETA models;

determine an ETA for a target route connecting the start location and the destination based on the global model; and

send the determined ETA to the user device.

2. The system of claim 1 , wherein to generate the global ETA model, the at least one processor is further directed to split each of the plurality of worker nodes into a root node and a plurality of leaf nodes based on a split rule.

3. The system of claim 2 , wherein to generate the global ETA model, the at least one processor is further directed to:

classify the plurality of leaf nodes into a plurality of groups; and

determine at least one barrier based on the plurality of leaf nodes, wherein the at least one barrier is configured between two of the plurality of groups of the leaf nodes.

4. The system of claim 3 , wherein the at least one barrier prevents the plurality of worker nodes from executing a first group of the leaf node and a second group of the leaf node simultaneously.

5. The system of claim 1 , wherein to allocate the feature data to the plurality of worker nodes, the at least one processor is further directed to:

determine a number of the plurality of worker nodes; and

allocate the feature data based on the number of the plurality of worker node on the plurality of worker nodes, wherein the feature data allocated on each of the plurality of worker nodes are different.

6. The system of claim 1 , wherein the plurality of worker nodes includes:

a first worker node operating a first sub-ETA model with first feature data as input;

a second worker node operating a second sub-ETA model with second feature data as input; and

the first worker node transmits the first feature data to the second worker node.

7. The system of claim 1 , wherein the global ETA model includes an Extreme Gradient Boosting model.

8. A method for determining an estimated time of arrival (ETA), comprising:

receiving, by at least one computer server via a network, a start location and a destination from a user device;

obtaining, by the at least one computer server, a global ETA model from a storage medium, wherein the global ETA model is generated according to a process, the process including:

obtaining feature data related to an on-demand service;

determining a plurality of sub-ETA models;

obtaining a parallel computing framework including a plurality of worker nodes, wherein each of the plurality of worker nodes is associated with a sub-ETA model of the plurality of sub-ETA models;

allocating the feature data to the plurality of worker nodes;

training the plurality of sub-ETA models based on the feature data; and

generating the global ETA model based on the plurality of trained sub-ETA models;

determining, by the at least one computer server, an ETA for a target route connecting the start location and the destination based on the global model; and

sending, by the at least one computer server via the network, the determined ETA to the user device.

9. The method of claim 8 , wherein the process of generating the global ETA model further includes splitting each of the plurality of worker nodes into a root node and a plurality of leaf nodes based on a split rule.

10. The method of claim 9 , wherein the process of generating the global ETA model further includes

classifying the plurality of leaf nodes into a plurality of groups; and

determining at least one barrier based on the plurality of leaf nodes, wherein the at least one barrier is configured between two of the plurality of groups of the leaf nodes.

11. The method of claim 10 , wherein the at least one barrier prevents the plurality of worker nodes from executing a first group of the leaf node and a second group of the leaf node simultaneously.

12. The method of claim 8 , wherein the allocating the feature data to the plurality of worker nodes includes:

determining a number of the plurality of worker nodes; and

allocating the feature data based on the number of the plurality of worker node on the plurality of worker nodes, wherein the feature data allocated on each of the plurality of worker nodes are different.

13. The method of claim 8 , wherein the plurality of worker nodes includes

a first worker node operating a first sub-ETA model with first feature data as input;

a second worker node operating a second sub-ETA model with second feature data as input; and

the first worker node transmits the first feature data to the second worker node.

14. The method of claim 8 , wherein the global ETA model includes an Extreme Gradient Boosting model.

15. A non-transitory computer readable medium, comprising at least one set of instructions for determining an estimated time of arrival (ETA), wherein when executed by at least one processor of an electronic terminal, the at least one set of instructions directs the at least one processor to perform acts of:

receiving a start location and a destination from a user device;

obtaining a global ETA model from a storage medium, wherein the global ETA model is generated according to a process, the process including:

obtaining feature data related to an on-demand service;

determining a plurality of sub-ETA models;

obtaining a parallel computing framework including a plurality of worker nodes, wherein each of the plurality of worker nodes is associated with a sub-ETA model of the plurality of sub-ETA models;

allocating the feature data to the plurality of worker nodes;

training the plurality of sub-ETA models based on the feature data; and

generating the global ETA model based on the plurality of trained sub-ETA models;

determining an ETA for a target route connecting the start location and the destination based on the global model; and

sending the determined ETA to the user device.

16. The non-transitory computer readable medium of claim 15 , wherein the generating the global ETA model further includes splitting each of the plurality of worker nodes into a root node and a plurality of leaf nodes based on a split rule.

17. The non-transitory computer readable medium of claim 16 , wherein the generating the global ETA model further includes:

classifying the plurality of leaf nodes into a plurality of groups; and

determining at least one barrier based on the plurality of leaf nodes, wherein the at least one barrier is configured between two of the plurality of groups of the leaf nodes.

18. The non-transitory computer readable medium of claim 17 , wherein the at least one barrier prevents the plurality of worker nodes from executing a first group of the leaf node and a second group of the leaf node simultaneously.