Data collation based on computer analysis of data
阅读说明:本技术 基于数据的计算机分析的数据核对 (Data collation based on computer analysis of data ) 是由 基兰·戈尔曼 克雷格·沃克 于 2018-07-03 设计创作,主要内容包括:提出了用于参照数据库中的数据来核对交易以基于针对交易提供的文本描述来识别交易参数的方法、系统和计算机程序。一种方法,包括:用于识别用于通过机器学习程序核对第一实体的交易的特征的操作。该特征包括至少对交易的描述、交易中第二实体的名称、第二实体的位置以及针对交易的账户。利用训练数据对机器学习程序进行训练,该训练数据包括用于先前核对的交易的特征的值。接收到的第一交易包括描述、日期和金额。第一交易被输入用于机器学习程序,该机器学习程序生成用于核对第一交易的一个或更多个建议。每个建议包括第一交易中的第二实体的名称和账户。(Methods, systems, and computer programs are presented for reconciling transactions with reference to data in a database to identify transaction parameters based on textual descriptions provided for the transactions. A method, comprising: an operation for identifying characteristics of a transaction for reconciling with a machine learning program a first entity. The characteristics include at least a description of the transaction, a name of the second entity in the transaction, a location of the second entity, and an account for the transaction. The machine learning program is trained using training data that includes values for features of previously collated transactions. The received first transaction includes a description, a date, and an amount. The first transaction is input for a machine learning program that generates one or more recommendations for reconciling the first transaction. Each suggestion includes a name and an account of the second entity in the first transaction.)
1. A method, comprising:
identifying characteristics of a transaction for reconciling a first entity by a machine learning program, the characteristics including a description of the transaction, a name of a second entity associated with the transaction, and an account associated with the transaction;
training, by one or more processors, the machine learning program with training data, the training data comprising values for features of previously collated transactions;
receiving, by the one or more processors, a first transaction, the first transaction comprising a description, a date, and an amount; and
inputting, by the one or more processors, the first transaction to the machine learning program, the machine learning program generating one or more suggestions for reconciling the first transaction, each suggestion including a name of the second entity in the first transaction and an account associated with the first transaction.
2. The method of claim 1, wherein the characteristics for reconciling transactions further comprise a location of the first entity, an industry of the first entity, and a location of the second entity.
3. The method of claim 1, wherein the characteristics for reconciling transactions further comprise a tax rate, an account identifier, an account name associated with the account identifier, an invoice identifier, and a billing identifier for the transaction.
4. The method of claim 1, wherein receiving the first transaction further comprises:
receiving a report having one or more transactions; and
extracting the one or more transactions, wherein each extracted transaction is verified using the machine learning program.
5. The method of claim 1, wherein the machine learning program identifies a score for each of the one or more suggestions, the method further comprising:
for the suggestion with the highest score, checking whether the score is above a first predetermined threshold; and
automatically checking the first transaction against the suggestion having the highest score when the highest score is above the first predetermined threshold.
6. The method of claim 5, further comprising:
checking if the highest score is above a second predetermined threshold when the highest score is not above the first predetermined threshold;
presenting a user interface for manually checking the first transaction against any of the one or more suggestions when the highest score is above the second predetermined threshold; and
requesting, in the user interface, a manual reconciliation for the first transaction without presenting any advice when the highest score is below or equal to the second predetermined threshold.
7. The method of claim 6, wherein the user interface includes a first field having information about the first transaction and a second field for the suggestion, the second field including a name of the second entity, an account, a billing identifier, a confirmation selector, and a manual selector for requesting a manual reconciliation.
8. The method of claim 1, wherein the training data is segmented for a plurality of geographic regions, wherein the machine learning program checks for transactions for one of the geographic regions.
9. The method of claim 1, wherein the training data comprises data obtained from a plurality of entities, wherein entities that do not have a history of verified transactions are provided with suggestions for verified transactions.
10. The method of claim 9, further comprising:
training a customized machine learning program with transactions previously reconciled by the first entity, the customized machine learning program providing a recommendation for the first entity based on transactions previously reconciled by the first entity;
generating, by the customized machine learning program, one or more customized suggestions for the first transaction; and
selecting a best suggestion to present to a user based on the suggestion and the customized suggestion.
11. A system, comprising:
a memory comprising instructions; and
one or more computer processors, wherein the instructions, when executed by the one or more computer processors, cause the one or more computer processors to perform operations comprising:
identifying characteristics of a transaction for reconciling a first entity by a machine learning program, the characteristics including a description of the transaction, a name of a second entity associated with the transaction, and an account associated with the transaction;
training the machine learning program with training data, the training data including values for features of previously collated transactions;
receiving a first transaction, the first transaction comprising a description, a date, and an amount; and
inputting the first transaction to the machine learning program, the machine learning program generating one or more suggestions for reconciling the first transaction, each suggestion including a name of the second entity in the first transaction and an account associated with the first transaction.
12. The system of claim 11, wherein the characteristics for reconciling transactions further comprise a location of the first entity, an industry of the first entity, a location of the second entity, a tax rate of the transaction, an account identifier, an account name associated with the account identifier, an invoice identifier, and a billing identifier.
13. The system of claim 11, wherein receiving the first transaction further comprises:
receiving a report having one or more transactions; and
extracting the one or more transactions, wherein each extracted transaction is verified using the machine learning program.
14. The system of claim 11, wherein the machine learning program identifies a score for each of the one or more suggestions, wherein the instructions further cause the one or more computer processors to perform operations comprising:
for the suggestion with the highest score, checking whether the score is above a first predetermined threshold; and is
Automatically checking the first transaction against the suggestion having the highest score when the highest score is above the first predetermined threshold.
15. The system of claim 14, wherein the instructions further cause the one or more computer processors to perform operations comprising:
checking if the highest score is above a second predetermined threshold when the highest score is not above the first predetermined threshold;
presenting a user interface for manually checking the first transaction against any of the one or more suggestions when the highest score is above the second predetermined threshold; and
requesting, in the user interface, a manual reconciliation for the first transaction without presenting any advice when the highest score is below or equal to the second predetermined threshold.
16. A non-transitory machine-readable storage medium comprising instructions that, when executed by a machine, cause the machine to perform operations comprising:
identifying characteristics of a transaction for reconciling a first entity by a machine learning program, the characteristics including a description of the transaction, a name of a second entity associated with the transaction, and an account associated with the transaction;
training the machine learning program with training data, the training data including values for features of previously collated transactions;
receiving a first transaction, the first transaction comprising a description, a date, and an amount; and
inputting the first transaction to the machine learning program, the machine learning program generating one or more suggestions for reconciling the first transaction, each suggestion including a name of the second entity in the first transaction and an account associated with the first transaction.
17. The machine-readable storage medium of claim 16, wherein the features for reconciling transactions further comprise a location of the first entity, an industry of the first entity, a location of the second entity, a tax rate of the transaction, an account identifier, an account name associated with the account identifier, an invoice identifier, and a billing identifier.
18. The machine-readable storage medium of claim 16, wherein receiving the first transaction further comprises:
receiving a report having one or more transactions; and
extracting the one or more transactions, wherein each extracted transaction is verified using the machine learning program.
19. The machine-readable storage medium of claim 16, wherein the machine learning program identifies a score for each of the one or more suggestions, wherein the machine further performs operations comprising:
for the suggestion with the highest score, checking whether the score is above a first predetermined threshold; and
automatically checking the first transaction against the suggestion having the highest score when the highest score is above the first predetermined threshold.
20. The machine-readable storage medium of claim 19, wherein the machine further performs operations comprising:
checking if the highest score is above a second predetermined threshold when the highest score is not above the first predetermined threshold;
presenting a user interface for manually checking the first transaction against any of the one or more suggestions when the highest score is above the second predetermined threshold; and
requesting, in the user interface, a manual reconciliation for the first transaction without presenting any advice when the highest score is below or equal to the second predetermined threshold.
Technical Field
The subject matter disclosed herein relates generally to methods, systems, and programs for reconciling two data sources based on a data description.
Background
Reconciliation is the process for confirming that an entry in the accounting system matches a corresponding entry in the bank statement. When the accounting receives the bank statement, the accounting must identify each entry in the bank statement to identify the corresponding account.
However, bank statements typically include ambiguous entries, which makes it difficult to identify corresponding accounts and parties. For example, an entry may not include the name of the payer, but rather provides a general description of the nature of the entry, such as tax, withdrawal, or payroll. Sometimes, an inference can be made of the name of the counterparty, such as by identifying a local entity that pays a property tax for the entry "property tax".
Bank reconciliation can be a difficult task due to the large degree of variability of the bank statement description, especially for computer programs that attempt to reconcile data automatically. A person may use his own experience to identify the nature of a transaction, but because of the lack of standards that provide a description for bank statements, automating a computer program to automatically identify the nature of a transaction and the parties to the transaction can be a difficult task.
Drawings
The various drawings in the figures illustrate only example embodiments of the disclosure and are not to be considered limiting of its scope.
Fig. 1 is a diagram illustrating transaction reconciliation according to some example embodiments.
Fig. 2A illustrates a process for reconciling transactions extracted from bank statements using a machine learning program, according to some example embodiments.
Fig. 2B is a user interface for reconciling transactions, according to some example embodiments.
Fig. 3 illustrates a process for training and using a machine learning program, according to some example embodiments.
Fig. 4 illustrates the use of two machine learning programs to verify a transaction, according to some example embodiments.
Fig. 5 is a flow chart of a method for reconciling transactions, according to some example embodiments.
FIG. 6 is a block diagram depicting an exemplary reconciliation platform according to some embodiments.
Fig. 7 is a flow diagram of a method for reconciling transactions with reference to data in a database to identify transaction parameters based on textual descriptions provided for the transactions, according to some example embodiments.
FIG. 8 is a block diagram depicting an exemplary application framework according to some embodiments.
Fig. 9 is a block diagram depicting an exemplary hosting architecture, in accordance with some embodiments.
FIG. 10 is a block diagram depicting an exemplary data center system for implementing embodiments.
FIG. 11 is a block diagram illustrating an example of a machine on which one or more example embodiments may be implemented.
Detailed Description
Example methods, systems, and computer programs relate to reconciling transactions with reference to data in a database to identify transaction parameters based on textual descriptions provided for the transactions. Examples merely typify possible variations. Unless explicitly stated otherwise, components and functions are optional and may be combined or subdivided, and the order of operations may vary or operations may be combined or subdivided. In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the example embodiments. It will be apparent, however, to one skilled in the art that the present subject matter may be practiced without these specific details.
The bank check has two aspects: payment and accountability, i.e., expenditure and income. For example, when an expense is identified in a bank statement, the reconciliation process must classify the expense as being associated with a person or label, such as a sale. In addition, the reconciliation process identifies another party in the transaction. Embodiments presented herein utilize data provided in bank report entries and access Machine Learning Programs (MLPs) trained with past reconciliation data to automatically reconcile report entries or provide good recommendations to accounting for reconciliation of entries.
MLP identifies the following verification features: such as information about the transaction (e.g., description, amount, date), information about the company (e.g., location, industry, number of employees), and accounting data (e.g., payee information, account number, account name, tax rate, invoice, bill), or any combination thereof. The system trains the MLP with the transactions that have been checked. Once trained, the MLP can receive entries from bank statements and provide suggestions on how to check the entries or automatically check the entries if the level of certainty estimated by the machine learning program is above a predetermined threshold.
This simplifies the work of checking accountants because accountants need only select from one or more recommendations for checking, or because the system automatically checks some entries without accounting doing the work.
In addition, the data used to train the MLP may be data collected from a large number of customers across geography. This means that new companies with little accounting history data can receive recommendations for reconciliation based on experience collected by the accounting service from other clients. In addition, the machine learning program may also utilize local models and understand the parameters and behavior of a particular company to generate good suggestions based on the company's accounting history.
One general aspect includes a method that includes operations for identifying characteristics for a transaction of a first entity to be collated by a machine learning program. The characteristics include a description of the transaction, a name of a second entity associated with the transaction, and an account associated with the transaction. The method also includes operations for training, by the one or more processors, a machine learning program with training data that includes values for features of previously collated transactions. The method further comprises the following operations: receiving, by one or more processors, a first transaction comprising a description, a date, and an amount; and inputting, by the one or more processors, the first transaction to the machine learning program. The machine learning program generates one or more suggestions for reconciling the first transaction, each suggestion including a name of the second entity in the first transaction and an account associated with the first transaction.
One general aspect includes a system that includes a memory having instructions and one or more computer processors. The instructions, when executed by one or more computer processors, cause the one or more computer processors to perform operations comprising: identifying characteristics of a transaction for reconciliation by a machine learning program of a first entity, the characteristics including a description of the transaction, a name of a second entity associated with the transaction, and an account associated with the transaction; training the machine learning program with training data, the training data including values for features of previously collated transactions; receiving a first transaction comprising a description, a date, and an amount; and inputting the first transaction to a machine learning program, the machine learning program generating one or more suggestions for reconciling the first transaction, each suggestion including a name of the second entity in the first transaction and an account associated with the first transaction.
One general aspect includes a non-transitory machine-readable storage medium containing instructions that, when executed by a machine, cause the machine to perform operations comprising: identifying characteristics of a transaction for reconciliation by a machine learning program of a first entity, the characteristics including a description of the transaction, a name of a second entity associated with the transaction, and an account associated with the transaction; training the machine learning program with training data, the training data including values for features of previously collated transactions; receiving a first transaction comprising a description, a date, and an amount; and inputting the first transaction to a machine learning program, the machine learning program generating one or more suggestions for reconciling the first transaction, each suggestion including a name of the second entity in the first transaction and an account associated with the first transaction.
Fig. 1 is a diagram illustrating transaction reconciliation according to some example embodiments. In accounting, reconciliation is the process of ensuring that two sets of records (e.g., balances of two accounts) are consistent. For example, a check is used to ensure that the money leaving the account matches the actual money spent. In addition, for bank accounts, reconciliation is the process used to confirm that the balance in the checkbook matches the corresponding bank statement. This includes matching entries in the bank statement with the account holder's payment or receipt.
In the example shown in fig. 1,
The payee remits the
During
Often, checking entries based on a short or ambiguous description in a bank statement is a challenge, which can make checking a tedious, tedious and boring task, where errors can occur. The goal of the accounting service is to make reconciliation an easy task (e.g., by providing suggestions to the user based on bank statements). For example, the amount may be a good indicator for generating a suggestion by matching the amount to an entry in the accounting system. However, an amount-based match does not always work because there may not yet be an entry in the accounting system, or because the payer may merge multiple payments into a single check.
Although sometimes the name of the payer may be included in the report, many times the name of the payee is not included in the report, but rather there is a description of a service, such as "taxi service" or "entertainment". These are some reasons why performing automatic reconciliation of bank statements in an accounting system may be difficult (and sometimes impossible) and requires manual reconciliation.
Some solutions to reconciliation are based on defining rules for reconciliation, e.g., "if the entry includes 'taxi', then the account is 2547 and a new accounting entry is added. However, the rules are difficult to collate across large volumes of reports. Rules are also inconvenient because someone must create and maintain the rules.
Embodiments presented herein are described with reference to checking received payments, but the same principles may be used to check made payments. The pattern for matching entries is different when reimbursing for payment or receiving payment, and the machine learning algorithm described below takes into account whether money is in or out when checking each entry.
Fig. 2A illustrates a process for reconciling transactions extracted from bank statements using a machine learning program, according to some example embodiments. The
When the system identifies the nature of an entry in a bank statement with a large degree of certainty, the embodiments presented herein facilitate reconciliation by providing suggestions to the user or by reconciling automatically without user intervention.
In some example embodiments, the
After cleaning and normalizing the description,
In some example embodiments, the
The
In some cases, a bank statement may check several accounting entries (e.g., payments of several invoices with one check) and present the user with multiple entries to indicate that one payment is related to several accounting entries.
Fig. 2B is a
Each transaction to be reconciled 224 includes one or more of: the date of the transaction (e.g., 3/2/2018), the name of the party in the transaction (e.g., ABC property management), reference information (e.g., a rent), and an amount (e.g., 1,181.25) that may be the amount spent or the amount received. In addition, the transaction to be reconciled 224 includes an option for deleting the transaction and an option for creating rules to handle this type of transaction. It should be noted that the system may also create rules over time based on past checks by the user, where a rule for a certain type of entry may be consistent with a certain account if the user performs the same check one or more times to match an entry in a bank statement with an account. Furthermore, in addition to the option to add rules, the user may also have options to modify or delete rules, even those automatically created by the system.
The proposed
A check button 228 (e.g., including the message "OK" to indicate an entry match) is provided as an option, and if the user selects
If the user prefers to create a new accounting entry for reconciliation, the create tab may be selected. The create tag may also be automatically presented if the accounting system does not find a suggestion to check the entry. Creating the label includes options for entering the name or party of the transaction, account, description, region, tax rate, and adding additional details for the transaction. After the user enters the information, a
The transfer label may be selected to mark when the transaction is the result of a transfer between bank accounts of the user, where both bank accounts are linked to the accounting system. If one of the bank accounts is not in the accounting system, a create operation may be used instead of the transfer.
Discussion tags allow a user to leave messages for other users and discuss reconciliation of transactions. For example, a user may enter "i don't know how to code this," while another user (e.g., accounting) may see the message and enter details for the transaction.
It should be noted that the embodiments shown in fig. 2A and 2B are exemplary and that not every possible embodiment is described. Other embodiments may utilize different layouts of the user interface, different fields, additional fields, present more than one suggestion at a time, and so forth. Accordingly, the embodiments illustrated in fig. 2A and 2B should not be construed as exclusive or limiting, but rather as illustrative.
Fig. 3 illustrates a process for training and using a machine learning program, according to some example embodiments. In some example embodiments, a machine learning program, also referred to as a machine learning algorithm or tool, is used to perform operations associated with the collation.
Machine learning is a field of research that gives computers the ability to learn without being explicitly programmed. The study and construction of machine learning exploration algorithms, also referred to herein as tools, that can learn from existing data and make predictions about new data. Such machine learning tools operate by building models from example training data 312 to represent data-driven predictions or decisions as outputs or assessments (e.g., reconciliation recommendations 320). Although example embodiments are presented with respect to some machine learning tools, the principles presented herein may be applied to other machine learning tools.
In various example embodiments, different machine learning tools are used. For example, the collation may be generated using polynomial naive Bayes (MNB), Support Vector Machines (SVM), polynomial Logistic Regression (LR), Random Forests (RF), Neural Networks (NN), matrix factorization, and other tools.
For example, MNB is one of the simpler models to build and use bayesian theorem to combine the observed data of the input features to produce a posterior probability. Training time is proportional to the number of samples given, and decision between classes typically requires selection of the label with the highest probability. The SVM is a discriminant model that provides linear separability in a higher dimension, and is a good model when the input features have a high dimension, but may be affected by the dimension if there are not enough samples. In addition, the SVM model uses one-to-one as a method of binarizing multi-class classifications, which means that the performance may be degraded as the number of samples increases.
In general, there are two types of problems in machine learning: classification problems and regression problems. The classification problem aims at classifying items into one of several categories (e.g., is the object apple or orange. Regression algorithms aim at quantifying some items, for example by providing a value that is a real number (e.g., what is the probability that the bank entry is consistent with a given payee). In some implementations, the example machine learning algorithm provides a reconciliation score (e.g., from 1 to 100) for the reconciliation suggestion matching the bank entry. The machine learning algorithm utilizes the training data 312 to find correlations between the recognized features 302 that affect the results.
In an example embodiment, the features 302 may be of different types and may include one or more of transaction features 304, company data features 306, and accounting data features 308. Other embodiments may utilize additional features as well.
The transaction characteristics 304 include data about the banking transaction such as a description, amount, and date of the transaction. The company data feature 306 includes information about the user's company, such as company name, company location, industry, number of employees, and the like. The accounting data feature 308 includes information about the accounting system and includes information about the payee (e.g., payee name, payee location or location, payee enterprise, payee address, payee contact name, etc.), account number, account name, tax rate, invoice, bill, etc.
In some example embodiments, training data 312 includes prior user checks, such as matches made manually or suggestions accepted by the user, to match bank entries to financial entries (e.g., cash receipts, check payments, transfers). In some example embodiments, the user checks for multiple client inputs by the accounting service, which means that the user account and name may be different for each client. However, in many cases, the accounting service utilizes standardized user accounts or user account names to be able to correlate information from different companies. In some example embodiments, account normalization is used to translate accounts from different users into a common standard.
In one example test implementation, 700000 entries are used for the training data, although other implementations may utilize a different number of entries, such as a number of one million or more.
In some example embodiments, the training data is converted into a vocabulary that may be used as input to the model. The vocabulary may include 100000 unique tokens or more and each token is associated with a word. In some example embodiments, for each word, a vector is created and the sentence is represented by a matrix of vectors that combine the words.
Using the training data 312 and the recognition features 302, the machine learning tool is trained at operation 314. The machine learning tool evaluates the values of the features 302 as the features 302 are correlated with the training data 312. The result of the training is a trained
When the
In some example embodiments, the training data 312 is divided by region, e.g., by country or county, and different machine learning programs are used to generate the collation for each region. Often, people spend most of their money on local businesses, so partitioning data by regions is a simple method to reduce the complexity of the program to speed up the operation of the machine learning program. Thus, one machine learning program may be used to generate a collation recommendation 320 for the United states, while another machine learning program may be used to generate a collation recommendation 320 for Australia.
In another example embodiment, the country in which the company is located is used as a feature and the machine learning program can generate a reconciliation recommendation based on the country in which the user is located, the identification of the country in which the transaction occurred, and the like.
In some example embodiments, the machine learning program utilizes only a limited set of the most common outputs (e.g., payees). In some example implementations, the top 100 most frequently used contexts (contexts) are used, and the reconciliation suggestion is provided only for the top 100 most frequently used contexts. In this way, the machine learning procedure can be faster, although this means that the trade-off is made by limiting the number of possible suggestions. In other example embodiments, a limited set is not enforced and all payees are considered for reconciliation.
In some example embodiments, a confidence threshold is utilized, and reconciliation recommendations having a reconciliation score above the confidence threshold are considered for automatic reconciliation or for presentation to a user. Those suggestions below the confidence threshold are discarded.
In other example embodiments, a threshold for automatic reconciliation is utilized, and those reconciliation recommendations having a reconciliation score higher than the threshold for automatic reconciliation are automatically reconciled without user authentication. For example, the threshold for automatic reconciliation may be set to 0.95, and if a reconciliation score of 0.97 is received, the bank entry is automatically reconciled due to the high confidence level provided by the machine learning program. In other example embodiments, automatic verification is not utilized and all recommendations must be confirmed by the user. In some cases, the automatic check may be activated or deactivated by the user. For example, the automatic reconciliation may be turned off for new users of the accounting service, and then turned on as more data is collected in the reconciliation confidence.
In fact, a large number of accounting entries are concentrated in a relatively small set of payees, typically a large organization. Given that these large organizations often appear on bank statements, a system can be built that identifies and displays these contacts. In particular, because of the sharing of data across multiple organizations, suggestions may be provided to new organizations in the financial service immediately upon their receipt of a bank statement.
Fig. 4 illustrates the use of two machine learning programs to verify a transaction, according to some example embodiments. In some example embodiments, the collation may be performed using a plurality of MLPs, each MLP being directed to a particular collation feature. For example, two machine learning programs are used to suggest a collation:
The
In some example embodiments, other models for utilizing different MLPs are possible, such as by having one MLP focus on businesses within a local region and having other MLPs focus on national or world models. The embodiment shown in fig. 4 is presented with reference to community predictors and user predictors, but the same principles can be applied to other MLP predictors based on geography. In example embodiments herein, MLP community predictors and MLP user predictors may also be combined with geographically oriented MLPs.
Personalized predictors are useful in large countries (e.g., the united states) because MLP user predictors can narrow down the list of a large number of possible payees in a large country to focus on areas where a company is conducting a large number of transactions. In some example embodiments, a national model is used in conjunction with a regional model to enable capture of national transactions as well as local transactions. In other example embodiments, the predictor may be decomposed by state or county.
In addition, data obtained using the
When a
At
At
At
It should be noted that the embodiment shown in fig. 4 is an example and that not every possible embodiment is described. Other embodiments may utilize a different MLP, additional MLPs, additional data for the MLP, and the like. Accordingly, the embodiments shown in fig. 4 should not be construed as exclusive or limiting, but rather illustrative.
Fig. 5 is a flow diagram of a
At
From
For each entry in the bank statement, a number of
At
At
In some example embodiments, the reconciliation system is tested by using some known reconciliation as input for the MLP to compare the extent to which the MLP suggests a match to the actual reconciliation data. For example, 70% of the collation data may be used for training MLPs, while 30% of the data is used for testing.
Without wishing to be bound by theory, in some example tests, the system showed that the suggested accuracy was about 90% for the most common payee. Furthermore, systems show that there is a tradeoff between speed, accuracy, and confidence. Depending on the amount of data used, the MLP generates recommendations faster or slower. Further, by adjusting the confidence level, suggestions with low probability of success may be filtered out. It should be noted, however, that even better results are desired for ordinary payees and other payees because the system utilizes learned experience and because additional computing resources are utilized.
FIG. 6 is a block diagram depicting an
Practice
The practice
An
The
Report
Core features 630 include components used by both accounting and organization. Accounting and
Reconciliation MLP316 enables generation of a reconciliation recommendation with a corresponding reconciliation score. Reconciliation MLP316 can interact with accounting and
The
The
Fig. 7 is a flow diagram of a
From
Further, from
Implementations may include one or more of the following features. In one example, the characteristics for reconciling the transactions further include a location of the first entity, an industry of the first entity, and a location of the second entity.
In one example, the characteristics used to reconcile the transaction also include a tax rate of the transaction, an account identifier, an account name associated with the account identifier, an invoice identifier, and a bill identifier.
In one example, receiving the first transaction further comprises receiving a statement having one or more transactions and extracting the one or more transactions, wherein each extracted transaction is collated with a machine learning program.
In another example, the machine learning program identifies a score for each of the one or more suggestions, wherein the
In one example, the
In one example, the user interface includes a first field having information about the first transaction and a second field for suggestion, the second field including a name of the second entity, an account, a billing identifier, a confirmation selector, and a manual selector for requesting a manual reconciliation.
In one example, the training data is segmented for a plurality of geographic regions, wherein the machine learning program checks for transactions for one of the geographic regions.
In one example, the training data includes data obtained from a plurality of entities, wherein entities that do not have a history of reconciliation transactions are provided with suggestions for reconciliation transactions.
In one example, the
FIG. 8 is a block diagram depicting an
The HTML and/or
Construct 840 provides a relay through which data is processed and presented to a user.
In some example embodiments, when MLP is called to analyze an entry (
Net 820,
The
Fig. 9 is a block diagram depicting an exemplary hosting architecture 900, according to some embodiments. The platform may be implemented using one or more containers (pods) 910. Each container 910 includes an application server Virtual Machine (VM)920 (shown in fig. 9 as application server virtual machines 920A-920C) specific to the container 910 and application server virtual machines (e.g., an internal service VM 930 and an application protocol interface VM 940) shared between the containers 910. Application server virtual machines 920-940 communicate with clients and third party applications via a network interface or API. Application server virtual machines 920-940 are monitored by application manager 950. In some example embodiments, application server virtual machines 920A-920C and API VM 940 are publicly accessible, while internal service VM 930 is not accessible by machines external to managed architecture 900. Application server VMs 920A through 920C may provide end-user services via an application or network interface. Internal service VM 930 may provide backend tools to application server VMs 920A-920C, monitor tools to application manager 950, or provide other internal services. API VM 940 may provide a programming interface to third parties. A third party may use the programming interface to build additional tools that rely on the features provided by the container 910.
Internal firewall 960 ensures that only permitted communications are allowed between database manager 970 and publicly accessible virtual machines 920-940. Database manager 970 monitors the primary SQL servers 980A and 980B and the redundant SQL servers 990A and 990B. Virtual machines 920-940 may be implemented using Windows8008R2, Windows 8012, or another operating system. The support server may be shared across multiple containers 910. The application manager 950, internal firewall 960, and database manager 970 may span multiple containers 910 within a data center.
FIG. 10 is a block diagram depicting an exemplary
Within each
A
FIG. 11 is a block diagram illustrating an example of a machine on which one or more example embodiments may be implemented. In alternative embodiments, the
As described herein, an example may include or be operated by logic or multiple components or mechanisms. A circuit device is a collection of circuits implemented in a tangible entity comprising hardware (e.g., simple circuits, gates, logic, etc.). The circuit device members may vary over time and potential hardware variability. The circuitry includes members that can perform specified operations, either alone or in combination, in operation. In an example, the hardware of the circuit arrangement may be invariably designed to perform specified operations (e.g. hardwired). In an example, the hardware of the circuit device may include variably connected physical components (e.g., execution units, transistors, simple circuits, etc.) to encode instructions specifying an operation, the variably connected physical components including a computer-readable medium that is physically modified (e.g., magnetically, electrically, movably disposed of invariant mass particles, etc.). When connecting physical components, the potential electrical properties of the hardware composition are changed, for example from an insulator to a conductor or from a conductor to an insulator. The instructions enable embedded hardware (e.g., an execution unit or a loading mechanism) to create members of a circuit device in the hardware via a variable connection to perform a portion of a specified operation when operating. Thus, when the apparatus is in operation, the computer readable medium is communicatively coupled to other components of the circuit arrangement. In an example, any of the physical components may be used in more than one member of more than one circuit device. For example, in operation, an execution unit may be used in a first circuit of a first circuit arrangement at one point in time and reused by a second circuit in the first circuit arrangement or by a third circuit in the second circuit arrangement at a different time.
The machine (e.g., computer system) 1100 may include a hardware processor 1102 (e.g., a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a hardware processor core, or any combination thereof), a
The
While the machine-
The term "machine-readable medium" may include any medium that is capable of storing, encoding or carrying
The
Throughout this specification, multiple instances may implement a component, an operation, or a structure described as a single instance. Although individual operations of one or more methods are illustrated and described as separate operations, one or more of the individual operations may be performed concurrently and nothing requires that the operations be performed in the order illustrated. Structures and functionality presented as discrete components in the example configurations may be implemented as a combined structure or component. Similarly, structures and functionality presented as a single component may be implemented as separate components. These and other variations, modifications, additions, and improvements fall within the scope of the subject matter herein.
The embodiments illustrated herein are described in sufficient detail to enable those skilled in the art to practice the disclosed teachings. Other embodiments may be utilized and derived from the description herein, such that structural and logical substitutions and changes may be made without departing from the scope of this disclosure. The detailed description is, therefore, not to be taken in a limiting sense, and the scope of various embodiments is defined only by the appended claims, along with the full range of equivalents to which such claims are entitled.
As used herein, the term "or" may be interpreted in an inclusive or exclusive sense. Further, multiple instances may be provided for a resource, operation, or structure described herein as a single instance. In addition, the boundaries between the various resources, operations, modules, tools (engines) and data stores are somewhat arbitrary, and particular operations are illustrated in the context of specific illustrative configurations. Other allocations of functionality are readily envisioned and may fall within the scope of various embodiments of the present disclosure. In general, structures and functionality presented as separate resources in the example configurations may be implemented as a combined structure or resource. Similarly, structures and functionality presented as a single resource may be implemented as separate resources. These and other variations, modifications, additions, and improvements fall within the scope of the embodiments of the disclosure as represented by the claims that follow. The specification and drawings are, accordingly, to be regarded in an illustrative rather than a restrictive sense.
- 上一篇:一种医用注射器针头装配设备
- 下一篇:生物认证结算系统、结算系统和收银系统