Intelligent price inquiry method and device for achieving spot trading

Abstract

The present invention discloses an inquiry method and device for intelligently achieving spot bond transactions, and relates to the field of artificial intelligence technology. The method includes: obtaining the spot bond trading tasks configured for the institution, and the configured risk verification rules; based on the spot bond trading tasks, automatically talking with potential transaction objects; obtaining the corpus content during the conversation with the potential transaction object, and when determining that the other party has a clear transaction intention, confirming the transaction elements during the conversation based on the corpus content; calling the risk verification module to perform risk verification on the transaction elements, and when the risk verification passes, collecting the complete transaction elements of the transaction and generating formatted transaction details of the intended transaction; sending the formal transaction to the foreign exchange trading center platform through the quotation sending module. The present invention improves the purposefulness of automatic conversation, helps to promote the intended transaction in the inter-bank spot bond market, and greatly reduces the manual workload while controlling transaction risks.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a price inquiry method and device for intelligently achieving spot bond transactions. Background Art

[0002] Bond trading forms include repurchase transactions, spot transactions, forward transactions and bond lending, etc. Among them, spot transactions refer to the transaction behavior in which the two parties transfer the ownership of bonds at an agreed price.

[0003] The interbank bond market is the dominant bond market in my country. As an over-the-counter market, spot transactions in the interbank bond market are completed through price inquiries, that is, traders of institutions participating in the interbank bond market inquire with their selected counterparties one by one through instant messaging software or applications such as QQ, telephone, interbank iDeal dialogue tools, dialogue quotes on the trading platform, and request quote windows. After reaching an agreement, the transaction is completed on the local currency trading platform of the foreign exchange center. This means that traders need to spend a long time manually talking to each potential counterparty, and spend a lot of time and energy to view, understand and record transaction elements, which affects transaction efficiency. At the same time, due to the large number of transaction elements involved in bond price inquiries, the large number of trading institutions, the rapid changes in quotation information, and the complex quotation message items, traders must constantly search for the information they are concerned about in a large number of quotation texts, and they also need to manually switch between different windows, which is not conducive to quick and accurate transaction value judgment, and it is difficult to meet the transaction requirements of strong competitiveness and high timeliness. Furthermore, in terms of risk control, the existing manual trading method cannot automatically complete the transaction risk check before setting the buying and selling tasks, responding to inquiries, and submitting the transaction order, which increases the market risk and operational risk faced by the trading institutions. Moreover, after the trading personnel and the counterparty reach a transaction intention in the later stage, they still need to manually enter the transaction quotation in the trading system of the foreign exchange trading center to finally complete the transaction, which consumes a lot of labor costs and time costs.

[0004] With the rapid development of artificial intelligence (AI), major financial institutions have begun to invest heavily in the research and development and application of artificial intelligence in the fields of trading, risk control, customer service, etc. AI uses natural language processing (NLP) and machine learning (ML) models to perform intelligent transaction data processing, which can make daily work more efficient and accurate, reduce labor consumption, and greatly meet the refined business demands of financial institutions.

[0005] As an example, Chinese patent application CN202310188187.1 provides a method for AI trader to inquire about prices and conduct intelligent transactions, including the following steps: S1: The AI ​​trader receives the other party's inquiry message and obtains the other party's identity; S2: The AI ​​trader initiates multiple rounds of inquiries to the other party to obtain the necessary transaction elements; S3: The AI ​​trader passes the necessary transaction elements to the foreign exchange market maker system, and the foreign exchange market maker system quotes and feeds back to the AI ​​trader and the other party; S4: If the other party cancels the transaction, the process ends; if the other party confirms the quotation, the counterparty is found and confirmed; S5: The AI ​​trader again initiates multiple rounds of inquiries to the other party to obtain the remaining non-essential transaction elements; S6: The AI ​​trader initiates a transaction according to the transaction method and obtains the corresponding information of the transaction method. The corresponding information of the transaction method is matched with all the information obtained by the AI ​​trader. If the match is successful, the transaction is completed. If the match is unsuccessful, manual intervention is performed. The above solution can automatically screen transaction intentions and match transactions by setting up AI traders (AI robots) to respond intelligently between the inquiring party and the foreign exchange market maker system, freeing up human resources and improving the accuracy and timeliness of foreign exchange inquiry quotations. In addition, the above solution uses ex ante credit risk control and operational risk control to enhance risk management capabilities: automatically query and pre-occupy the counterparty's credit line before the transaction to conduct ex ante credit risk control, automatically check the transaction intention and the actual order transaction elements during the transaction, assist human traders to complete transaction confirmation, and realize operational risk control.

[0006] However, the above technical solution has the following defects when applied to bond trading: First, although AI traders can respond and quote intelligently, they need to respond only when they receive the other party's inquiry message. The communication behavior is relatively passive, and the communication content is based on the other party's inquiry information, and the transaction characteristics of the institution are not taken into account. Secondly, the control of operational risks by this solution is limited to checking the consistency of the transaction elements of the transaction intention and the actual order, and it is unable to conduct risk control on the consistency and deviation between the other party's quotation, the party's reply quotation and the party's buying and selling tasks, and the market situation. Summary of the invention

[0007] The purpose of the present invention is to overcome the shortcomings of the prior art and provide an inquiry method and device for intelligently achieving spot transactions. The present invention can conduct targeted conversations based on the configuration tasks of the institution during the automatic conversation stage, automatically initiate inquiries or automatically match received inquiries based on the actual buying and selling tasks required by the institution, thereby improving the purposefulness of the automatic conversation and facilitating the intended transaction in the interbank spot market; at the same time, after confirming the transaction elements, the transaction intention is risk checked based on the risk check rules of the institution, which is based on a preset quotation strategy. After the risk check passes, the complete transaction element details of the transaction are collected and the formatted transaction details of the intended transaction are generated, and then the formal transaction is sent to the foreign exchange trading center platform through the quotation sending module, which controls the transaction risks while greatly reducing the manual workload.

[0008] To achieve the above objectives, the present invention provides the following technical solutions:

[0009] A method for intelligently completing a spot transaction includes the following steps:

[0010] Obtain the current securities trading tasks configured for the institution, as well as the configured risk verification rules;

[0011] Based on the above-configured current bond trading task, an automatic conversation is conducted with a potential trading partner regarding the current bond trading intention; the automatic conversation includes automatically initiating a price inquiry to the potential trading partner according to the above-configured current bond trading task, or automatically matching the price inquiry information with the above-configured current bond trading task after receiving the price inquiry information from the potential trading partner, and automatically replying to the other party when the price inquiry information can match the above-configured current bond trading task;

[0012] Obtaining the content of the corpus during the conversation with the potential transaction partner, judging whether the other party has a clear transaction intention based on the content of the corpus, and when judging that the other party has a clear transaction intention, confirming the transaction elements in the aforementioned conversation process based on the content of the corpus;

[0013] Based on the risk check rules configured above, the risk check module performs risk check on the transaction intention; when the risk check passes, it is determined that the transaction intention is achieved. At this time, the complete transaction element details of the transaction are collected and formatted transaction details of the intended transaction are generated;

[0014] According to the formatted transaction details of the aforementioned intended transaction, the formal transaction is sent to the foreign exchange trading center platform through the quotation sending module.

[0015] Furthermore, the current bond trading task is configured according to the transaction preference of the institution, and the configured current bond trading task is a single task or a batch task;

[0016] One or more transaction preference elements of the institution are recorded in the current bond trading task; the transaction preference elements include transaction direction, transaction type, bond code, settlement speed, target volume, single transaction volume, validity period, price, whether to send price and / or whether to reply to the other party's inquiry.

[0017] Further, when confirming the transaction elements in the aforementioned conversation process based on the corpus content, a preset transaction data recognition model is called to perform natural language analysis on the corpus content to identify the transaction element information in the corpus;

[0018] The transaction data recognition model adopts a combination of the BERT-CRF model and transaction rules when identifying transaction elements. The transaction rules include the transaction rules of the interbank bond market. The type, quantity and / or format of the transaction elements extracted by the transaction data recognition model meet the requirements of the transaction rules.

[0019] Furthermore, the steps of judging whether the other party has a clear transaction intention and confirming the transaction elements according to the content of the corpus include:

[0020] When this party initiates an inquiry, the transaction elements of the inquiry sent to the potential transaction partner include at least the bond code, transaction volume and settlement speed; at this time, the corpus content of the conversation between this party and the potential transaction partner is identified to determine whether the corpus contains content that the potential transaction partner expects to conclude a transaction and / or quotation content with clear transaction elements. If the above content is included, it is determined that the potential transaction partner has a clear transaction intention; at this time, the steps of confirming the transaction elements are as follows: this party sends a request to the potential transaction partner to confirm the relevant transaction elements, and after receiving the potential transaction partner's confirmation of the above request, the above transaction elements are confirmed;

[0021] When we receive an inquiry from a potential transaction partner, we determine whether the bond code and transaction direction of the inquiry match the aforementioned current bond buying and selling tasks, and if they match, we determine that the potential transaction partner has a clear transaction intention; at this time, the steps to confirm the transaction elements are as follows: we send a request to the potential transaction partner to supplement other relevant transaction elements, and after receiving the supplementary transaction element information from the potential transaction partner, we integrate the supplemented transaction elements with the existing transaction elements and then confirm the transaction elements.

[0022] Furthermore, corresponding to the method of receiving the inquiry information of the potential transaction object, the step of automatically replying to the other party includes:

[0023] Collect the identity information of the potential transaction object and determine whether the potential transaction object is on the blacklist preset by the institution. If it is determined that the potential transaction object is on the blacklist, the automatic conversation process ends; otherwise, proceed to the next step;

[0024] After collecting the bond code and transaction direction inquired by the potential transaction object, match them with the aforementioned current bond trading task. If it is determined that it cannot be matched with the aforementioned current bond trading task, it will prompt the potential transaction object that there is no quotation from this party; otherwise, proceed to the next step;

[0025] Collect the price and transaction volume information provided by the potential transaction partner when inquiring, and determine whether the provided price and transaction volume information is unfavorable to the party. If it is determined that the provided price or transaction volume is unfavorable to the party, prompt the potential transaction partner that the transaction cannot be concluded; otherwise, proceed to the next step;

[0026] We will automatically reply to the above inquiry, and the automatic reply information will include our quotation information.

[0027] Further, when the risk check module is used to check the risk of the aforementioned transaction intention, the risk check is performed according to the quotation strategy configured by the party;

[0028] At this time, a price strategy setting module is configured to configure the quotation strategy of the party; the market price benchmark in the quotation strategy selects the quotation quotation, and accepts the BP deviation setting of the Bid Spread / Ofr Spread based on the quotation quotation. After setting, the recommended price is automatically obtained when quoting;

[0029] According to the quotation strategy, the risk check module is configured as follows: when the other party's quoted price is better than the price set by our party, the check is judged to be passed, our party accepts the quotation and proceeds to the next process; when the other party's required quantity is greater than the target quantity or the quoted price is unfavorable to our party, the check is judged to be failed, a reminder that the transaction intention cannot be reached is issued, and when the preset conditions are met, the reason why the transaction intention cannot be reached is sent to the other party.

[0030] Furthermore, when automatically replying to the price inquiry of the potential transaction object, it also includes the step of calling the risk check module to perform risk check on the replied price information;

[0031] The check items configured in the risk check module include a price deviation check item, a price deviation value check item, a quotation quantity check item and a transaction volume check item; at this time, the risk check module is configured to: collect the quotation information in the reply information of the party, determine whether the aforementioned quotation information meets the limit range of the aforementioned check items, when the quotation information meets the limit range of the aforementioned check items, determine that the check is passed, and send the reply information of the party to the aforementioned potential transaction object; when the quotation information does not meet the limit range of one or more of the aforementioned check items, determine that the check is failed, prompt the potential transaction object that the party has no quotation yet, and terminate the transaction process.

[0032] Furthermore, the formatted transaction details are a formatted transaction intention form to be executed, and the formatted transaction intention form is structured data using a preset data format, and includes necessary transaction elements required under the inter-bank bond market transaction rules.

[0033] Furthermore, it includes a dialogue quotation mode and a quotation request mode. According to different quotation modes, the quotation sending module is configured as follows:

[0034] In the dialogue quotation mode, when the formatted transaction intention form is initiated by this party, the aforementioned formatted transaction details are collected and the quotation for the formal transaction is sent to the foreign exchange trading center platform; when the formatted transaction intention form is initiated by the potential trading counterparty, after the party identifies the source of the received quotation information, it confirms the formal transaction to the foreign exchange trading center platform;

[0035] In the quotation request mode, when we are the party requesting the quotation, we initiate the sending of the formatted transaction intention form, collect the aforementioned formatted transaction details, and send the formal transaction to the foreign exchange trading center platform; when we are the party requesting the quotation, the formatted transaction intention form can be converted into a proposed quotation order. At this time, after identifying the quotation information received by us, we match the proposed quotation order with the front-end transaction according to the elements. After the match is successful, we reply to the relevant elements to the inquirer. After the inquirer agrees to the transaction, we confirm the formal transaction with the foreign exchange trading center.

[0036] Furthermore, during the entire transaction process, reminders are sent to potential transaction partners according to preset time limits and / or reminder scenarios, and the latest dynamic information of the transaction process is displayed.

[0037] The present invention also provides a price inquiry device for intelligently achieving spot trading, comprising the following structure:

[0038] Initialization module, used to obtain the current securities trading tasks configured for the institution, as well as the configured risk verification rules;

[0039] A dialogue module is used to automatically communicate with potential trading partners about their intention to trade current securities based on the aforementioned configured current securities trading tasks; the automatic conversation includes automatically initiating a price inquiry to the potential trading partner based on the aforementioned current securities trading tasks, or automatically matching the price inquiry information with the aforementioned configured current securities trading tasks after receiving the price inquiry information from the potential trading partner, and automatically replying to the other party when the price inquiry information can match the aforementioned current securities trading tasks;

[0040] The corpus analysis module is used to obtain the corpus content during the conversation with the potential transaction partner, and determine whether the other party has a clear transaction intention based on the corpus content. When it is determined that the other party has a clear transaction intention, the transaction elements in the aforementioned conversation process are confirmed based on the corpus content;

[0041] A risk check module, used to perform risk check on the transaction intention based on the risk check rules configured above;

[0042] A formatted transaction details generation module is used to determine that the aforementioned transaction intention has been reached when the risk check has passed. At this time, the complete transaction element details of the transaction are collected and the formatted transaction details of the intended transaction are generated;

[0043] The quotation sending module is used to send the formal transaction to the foreign exchange trading center platform according to the formatted transaction details of the aforementioned intended transaction.

[0044] Due to the adoption of the above technical solution, the present invention has the following advantages and positive effects compared with the prior art, as an example:

[0045] On the one hand, the present invention can conduct targeted conversations based on the configuration tasks of the own institution during the automatic conversation stage, automatically initiate inquiries or automatically match received inquiries based on the buying and selling tasks actually required by the own institution, thereby improving the purposefulness of automatic conversations and helping to facilitate intended transactions in the interbank spot bond market; at the same time, after confirming the transaction elements, the transaction intention is risk-checked based on the risk check rules of the own institution, and the risk check rules are based on a preset quotation strategy. After the risk check is passed, the complete transaction element details of the transaction are collected and a formatted transaction detail of the intended transaction is generated, and then the formal transaction is sent to the foreign exchange trading center platform through the quotation sending module, thereby managing transaction risks while greatly reducing manual workload.

[0046] On the other hand, according to real-time market data and historical transaction records, the current securities trading tasks and risk verification rules that meet the actual needs of the party can be configured. When configuring the current securities trading tasks of the day for the institution, it can be set as a single task or a batch task (a combination of multiple tasks) according to actual needs. The current securities trading tasks can record one or more transaction preference elements of the institution, such as transaction direction, transaction type (multiple securities / single securities), bond code, clearing speed, target volume, single transaction volume, effective time, price (maturity yield, net price, quotation strategy), whether to send the price, whether to reply to the other party's inquiry, and other element information. The risk verification rules can set different strategies and / or verification items according to different stages of the transaction, and perform relevant risk verification before, during, and after the automatic conversation. For example, when adding a new current bond trading task, after obtaining the configured current bond trading task, the task can be checked according to the pre-configured check items - for example, the check items can include the quotation amount, price deviation, price deviation value, etc. After the check is passed, a message can be sent to the institution to confirm whether the task is confirmed. After receiving the confirmation operation returned by the institution, it is determined that the current bond trading task has been successfully added, and the subsequent automatic conversation process can be entered; if the check fails, a reminder can be sent to the party, prompting that the task cannot be added or the added task can be directly rejected. For another example, during the automatic conversation process, if the other party inquires about the price of the party, the risk control check can also be performed when the party automatically replies to the price. The check items can include price deviation and price deviation value, etc. If the risk control check fails, the robot will prompt the other party that the party has no quotation for the time being, and the conversation ends. For another example, after the other party has made clear the transaction intention, after confirming the transaction elements in the chat process according to the corpus, the risk check can also be performed according to the configured quotation strategy.

[0047] On the other hand, depending on the quotation method, it is possible to initiate or identify the quotation source and send the formal transaction to the foreign exchange trading center platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flow chart of a method for intelligently inquiring about spot transactions provided in an embodiment of the present invention.

[0049] Figure 2 An information transmission diagram of the price inquiry method provided in an embodiment of the present invention.

[0050] Figure 3 A module structure diagram of an intelligent device for inquiring about spot transactions provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The following is a further detailed description of the price inquiry method and device for intelligent spot trading disclosed in the present invention in conjunction with the accompanying drawings and specific embodiments. It should be noted that the technologies (including methods and devices) known to ordinary technicians in the relevant fields may not be discussed in detail, but in appropriate cases, the above-mentioned known technologies are considered to be part of the specification. At the same time, other examples of exemplary embodiments may have different values. The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the invention.

[0052] In the description of the embodiments of the present application, " / " means or, and "and / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" means: A and B exist alone, B exists alone, and A and B exist at the same time. In the description of the embodiments of the present application, "multiple" means two or more.

[0053] Terminology explanation:

[0054] (1) ComStar system is a transaction management system for interbank market transaction data management. ComStar system consists of multiple functional modules. Through the combination of different functional modules, it supports different business lines and builds a management system for banks from information acquisition, front-end transaction analysis and management, middle-end risk control to back-end transaction direct processing.

[0055] (2) BERT-CRF model: The BERT-CRF model is a sequence labeling model that combines the pre-trained language model BERT (Bidirectional Encoder Representations from Transformers) and conditional random fields (Conditional Random Field), and is commonly used in natural language processing. BERT is a pre-trained model based on Transformer, which can effectively extract and process semantic information in text; CRF is a model based on conditional random fields, which can consider the features in text in an orderly manner, thereby achieving accurate recognition of new entities. The BERT-CRF model is essentially a CRF model, but it uses the BERT model to train the emission matrix in the CRF model. It extracts features from text data through the BERT model, and then inputs the extracted features into the CRF model for recognition and classification of new entities.

[0056] The technical concept and solution of the present invention are introduced below based on exemplary application scenarios. Example

[0057] The present invention improves the prior art described in the background technology. This embodiment provides an inquiry method for intelligently achieving spot trading. The method can be applied to the ComStar system. Figure 1 As shown, the method comprises the following steps:

[0058] S100, obtaining the current securities trading tasks configured for the institution and the configured risk verification rules.

[0059] Specifically, the current bond buying and selling task is configured according to the transaction preference of the institution, and the configured current bond buying and selling task is a single task or a batch task, and the batch task is a combination of multiple tasks.

[0060] The trading preference refers to the trading tendency of the party for different bonds based on its own trading habits, trading strategies, asset allocation requirements, real-time market conditions or historical trading data, etc. In this embodiment, the trading preference can be reflected by specific trading preference elements, which include but are not limited to trading direction, trading type, bond code, settlement speed, target volume, single transaction volume, effective time, price, whether to send price, whether to reply to the other party's inquiry, etc. When configuring the current bond trading task for the day for the party's institution, one or more trading preference elements of the party's institution can be recorded in the current bond trading task.

[0061] The current bond buying and selling tasks formulated according to the party's trading preferences are essentially trading tasks that concretize the party's trading preferences in order to achieve the party's trading objectives. For example, a configured buying and selling task information can be a current bond with a bond code of 220220, a target volume of 300 million, a transaction direction of bid (buy), and a quoted yield to maturity of 2.2%.

[0062] Optionally, the current bond trading task can be added, deleted or modified according to the needs of the party.

[0063] In the step S100, risk check rules are also configured for the institution. In specific settings, the institution can configure the check strategies and / or check items at different stages of the transaction process according to the needs of risk management and control, such as respectively configuring corresponding risk check rules for the three stages before the automatic conversation, during the automatic conversation, and after the automatic conversation, including a first check rule for risk check before the conversation, a second check rule for risk check during the conversation, and a third check rule for risk check after the conversation.

[0064] As an example of a typical method, if the user turns on the risk check before the automatic conversation (corresponding to the first check rule), the check items configured in the first check rule will be checked before adding or modifying the current bond trading task. Preferably, the check items supported by the first check rule may include transaction volume, price deviation and / or price deviation value.

[0065] The specific verification steps may be as follows: obtaining the transaction elements of the newly added or modified current securities trading task, such as the trading volume and price; comparing the obtained trading volume of the trading task with the trading volume in the preset risk verification rules. If the trading volume of the trading task does not meet the trading volume requirements in the risk verification rules, the risk verification fails. As an example and not a limitation, when the obtained buying task trading volume is greater than the trading volume in the risk verification rules, the risk verification fails.

[0066] The price of the acquired buying and selling tasks is calculated in combination with real-time market data, historical data, forecast data or other preset benchmark values ​​(i.e. preset reference prices) to calculate the price deviation and / or price deviation value. When the price deviation and / or price deviation value does not meet the threshold value specified in the risk check rules, the risk check fails.

[0067] Furthermore, some check items can be set as acceptable check items and mandatory rejection check items as needed. For the acceptable check items, if the trading task does not meet the requirements of the check items, a prompt of failure to pass the check will be triggered, but the party can forcibly add or modify the current bond trading task as needed; for the mandatory rejection check items, if the trading task does not meet the requirements of the check items, the current bond trading task cannot be added or modified.

[0068] S200, based on the above-configured current bond trading task, automatically communicate with potential trading partners about current bond trading intentions.

[0069] The potential transaction object refers to the transaction object with which a transaction may be concluded. The potential transaction object may be a trader who is talking to a chat tool to conclude a spot bond transaction, may be all transaction parties that conduct spot bond transactions on the trading platform, or may be a bond transaction object that is screened from all objects that conduct spot bond transactions on the trading platform and can provide a match with the transaction preference elements in the aforementioned spot bond transaction task.

[0070] The automatic conversation includes automatically initiating a price inquiry to a potential transaction partner according to the aforementioned current bond trading task, or automatically chatting after receiving price inquiry information from a potential transaction partner.

[0071] When the automatic conversation automatically initiates a price inquiry to a potential trading partner based on the aforementioned current bond trading task, the transaction elements of the price inquiry sent to the potential trading partner include at least the bond code, transaction volume and settlement speed, and the quotation can be sent or not as needed.

[0072] Preferably, when the automatic conversation is a conversation after receiving inquiry information from a potential transaction partner, the following processing steps (1)-(5) are preferably adopted.

[0073] Step (1), if the party has set up a blacklist in the current securities transaction, then after receiving the inquiry information of the potential transaction object, the identity information of the potential transaction object is collected and based on the blacklist preset by the party’s institution, it is determined whether the potential transaction object is in the aforementioned blacklist. If it is determined that the potential transaction object is in the aforementioned blacklist, the automatic conversation process is terminated.

[0074] Step (2), when it is determined that the potential transaction partner is not included in the blacklist, the bond code and transaction direction inquired by the potential transaction partner are collected and matched with the aforementioned current bond trading task. When it is determined that it cannot be matched with the aforementioned current bond trading task, the potential transaction partner is prompted that there is no quotation at present.

[0075] Step (3), after determining that the bond code and transaction direction inquired by the potential transaction object match the aforementioned current bond buying and selling task, collect the price and transaction volume information provided by the potential transaction object when inquiring, determine whether the provided price and transaction volume information is unfavorable to the party, and if it is determined that any of the provided price or transaction volume is unfavorable to the party, then prompt the potential transaction object that the transaction cannot be concluded. As an example but not a limitation, for example, the purchase task of the party is a current bond with a bond code of 220220, a price of a yield to maturity of 2.2%, and a transaction volume of 300 million yuan. When the potential transaction object quotes 2.1%, the quote will be determined to be unfavorable to the party, or if the potential transaction object intends to sell a transaction volume of more than 300 million yuan, the transaction volume will be determined to be unfavorable to the party. At this time, because the price or transaction volume is unfavorable to the party, it will be prompted that the potential transaction object cannot conclude the transaction.

[0076] Step (4), when it is determined that the price or transaction volume provided by the potential transaction partner meets the trading task of the party or is more beneficial to the party than the trading task, it is determined that the quotation of the party can be replied to the potential transaction partner, and an automatic / manual confirmation step can be configured during the process to indicate that the party accepts the price. More preferably, when automatically replying to the inquiry of the potential transaction partner, before sending the automatic reply information to the potential transaction partner, it can also include the step of calling the risk check module to perform a risk check on the replied price information. The check items configured in the risk check module can include a price deviation check item, a price deviation value check item, a quotation amount check item, and a transaction volume check item. At this time, the risk verification module is configured to: collect the quotation information in the reply information of the party, determine whether the aforementioned quotation information meets the limit range of the aforementioned verification items (that is, meets the requirements of all verification items); when the quotation information meets the limit range of the aforementioned verification items, determine that the verification is passed, and send the reply information of the party to the aforementioned potential transaction object; when the quotation information does not meet the limit range of one or more of the aforementioned verification items, determine that the verification fails, prompt the potential transaction object that the party has no quotation yet, and terminate the transaction process.

[0077] In this embodiment, the check item can be set with relevant limit information, and the check is judged to be passed when it is within the limit range, and the check is judged to be failed when it is outside the limit range. Taking the transaction volume as an example, for example, a lower limit can be set for the transaction volume, and the check is judged to be failed when it is lower than the lower limit.

[0078] Step (5), when no risk check step is set in the previous step, or the quotation replied by the party passes the set risk check step, an automatic reply is made to the aforementioned inquiry, the automatic reply information includes the quotation information of the party, and the automatic reply is sent to the potential transaction object.

[0079] S300, obtaining the content of the corpus during the conversation with the potential transaction partner, judging whether the other party has a clear transaction intention based on the content of the corpus, and when it is judged that the other party has a clear transaction intention, confirming the transaction elements in the aforementioned conversation process based on the content of the corpus.

[0080] Specifically, when we initiate an inquiry, the transaction elements of the inquiry sent to the potential transaction partner include at least the bond code, transaction volume and settlement speed. At this time, we identify the content of the conversation between us and the potential transaction partner, and determine whether the corpus contains content that the potential transaction partner expects to conclude a transaction and / or quotation content with clear transaction elements. If it contains the above content, we determine that the potential transaction partner has a clear transaction intention. As an example and not a limitation, if we have attached key information such as quotations when inquiring, and get an affirmative answer from the other party, such as "deal", then we determine that the other party has a clear transaction intention; for another example, if we do not attach a quotation when inquiring, or bargain with the potential transaction partner with the attached quotation, so that the party replies to us with quotation content containing clear transaction elements such as price, transaction volume and / or settlement speed, then we determine that the other party has a clear transaction intention.

[0081] At this time, the steps for confirming the transaction elements are as follows: the party sends a request for confirming the relevant transaction elements to the potential transaction object, and after receiving the confirmation of the potential transaction object for the aforementioned request, confirms the aforementioned transaction elements.

[0082] When we receive an inquiry from a potential trading partner, we determine whether the bond code and trading direction of the inquiry match the aforementioned current bond buying and selling task, and if there is a match, we determine that the potential trading partner has a clear trading intention.

[0083] At this time, the steps for confirming the transaction elements are as follows: this party sends a request to the potential transaction object to supplement other relevant transaction elements. After receiving the supplementary information of the transaction elements from the potential transaction object, the supplemented transaction elements are integrated with the existing transaction elements and then the transaction elements are confirmed.

[0084] In this embodiment, in the process of identifying and confirming transaction elements, a preset transaction data recognition model can be called.

[0085] Preferably, the aforementioned transaction data recognition model uses the BERT-CRF model when identifying transaction elements. The latest words of broker transaction data are trained by BERT-CRF to be close to the real transaction business scenario. The trained BERT-CRF model is used to identify transaction element information in natural language text to improve the accuracy of extracting elements. The BERT-CRF model can eliminate interference factors based on contextual semantic information. The specific BERT-CRF model training process and entity (transaction element) recognition method refer to the prior art and will not be repeated here.

[0086] At the same time, in order to further improve the accuracy of transaction element identification, the transaction data identification model is also combined with transaction rules, including the transaction rules of the inter-bank bond market. This requires that the type, quantity and / or format of the transaction elements extracted by the transaction data identification model must comply with the requirements of the transaction rules.

[0087] The transaction data recognition model mentioned above is used to perform natural language analysis on the corpus content, thereby identifying the transaction element information in the corpus.

[0088] S400, based on the risk check rules configured above, the transaction elements included in the transaction intention are risk checked by the risk check module; when the risk check passes, it is determined that the transaction intention is achieved. At this time, the complete transaction element details of the transaction are collected and the formatted transaction details of the intended transaction are generated.

[0089] In this embodiment, the risk check at this stage is performed according to the quotation strategy configured by the party. At this time, a special price strategy setting module can be configured to configure the quotation strategy of the party, and the price strategy setting module can be configured in the ComStar system. The market price benchmark in the quotation strategy selects the quotation quotation - for example, you can select the interbank market-making quotation, the interbank indicative quotation, the interbank X-Bond quotation, the market-making quotation of the party, the Xbond quotation of the party, the valuation data (CDC), etc., and accept the BP (basis point) deviation setting of the Bid Spread / Ofr Spread based on the quotation quotation, and accept the BP deviation setting of the Bid Spread / Ofr Spread based on the quotation quotation. After setting, the recommended price is automatically obtained when quoting. Preferably, the recommended price is a related price setting based on the quotation quotation, such as directly obtaining the CDC valuation or other valuations, or obtaining the related deviation setting of the buy / sell plus BP (basis point) based on the CDC valuation.

[0090] According to the quotation strategy, the risk check module is configured as follows: when the other party's quoted price is better than the price set by the party, the check is determined to be passed, the party accepts the quotation and proceeds to the next process; when the other party's required quantity is greater than the target quantity or the quoted price is unfavorable to the party, the check is determined to be unsuccessful, a reminder that the transaction intention cannot be reached is issued, and, when the preset conditions are met, the reason why the transaction intention cannot be reached is sent to the other party. In other words, at this time, the system can send the specific reason why the transaction intention cannot be reached to the counterparty according to the preset conditions, such as sending the maximum transaction volume information to the other party, and further, it can also give specific price suggestions, such as the best price.

[0091] In this embodiment, the risk check items configured at this stage may include transaction volume, price deviation and / or price deviation value.

[0092] If the risk check fails, a reminder will be triggered that the relevant intention cannot be achieved; if the risk check passes, it will be determined that the aforementioned transaction intention is achieved. At this time, the complete transaction element details of the transaction are collected and the formatted transaction details of the intended transaction are generated.

[0093] The formatted transaction details are a formatted transaction intention form to be executed. The formatted transaction intention form is structured data in a preset data format and includes necessary transaction elements required under the inter-bank bond market transaction rules.

[0094] The necessary transaction elements are those that can comprehensively and accurately summarize the transaction subject, transaction conditions and transaction methods of both parties and meet the requirements of the transaction system, and are indispensable for completing the transaction. By way of example and not limitation, the necessary transaction elements may include transaction date, investment portfolio, bond code, settlement speed, own direction, net price, yield to maturity, transaction volume, counterparty institution, counterparty trader, dialogue institution, dialogue trader, order issuance method, transaction source, dialogue status and / or local currency status. It should be noted that the necessary transaction elements may change due to system updates, changes or innovations in transaction processes.

[0095] S500, sending the formatted transaction details obtained in the previous step to the foreign exchange trading center platform.

[0096] Among them, according to the different quotation methods, the sending method can be divided into two situations: dialogue quotation and request quotation, see Figure 2 shown.

[0097] Specifically, under the dialogue quotation mode, when the formatted transaction intention form is initiated by the party, the aforementioned formatted transaction details are collected, and the quotation for the formal transaction is sent to the foreign exchange trading center platform; when the confirmation information of the potential transaction counterpart for the aforementioned quotation is collected, the transaction is completed. When the formatted transaction intention form is initiated by the potential transaction counterpart, the party identifies the source of the received quotation information and confirms the formal transaction to the foreign exchange trading center platform.

[0098] Under the quotation request method, there are also two situations: one is that this party is the inquirer requesting a quotation. At this time, this party, as the inquirer, actively initiates a quotation to the market maker; the other is that this party is the respondent requesting a quotation. At this time, the formatted transaction intention order can be converted into a proposed response order.

[0099] In specific implementation, when we are the party requesting quotation, we will initiate the sending of formatted transaction intention form, collect the aforementioned formatted transaction details, and send the formal transaction to the foreign exchange trading center platform.

[0100] When we are the one who requests a quotation, the formatted transaction intention form can be converted into a proposed quotation order. At this time, after identifying the quotation information received by us, the proposed quotation order can be matched with the front-end transaction according to the elements, and the relevant elements will be returned to the inquiring party after the match is successful. After the inquiring party agrees to the transaction, the formal transaction confirmation is submitted to the foreign exchange trading center.

[0101] In another implementation of this embodiment, during the entire transaction process, reminders can be sent to potential transaction partners according to preset time limits and / or reminder scenarios, and the latest dynamic information of the transaction process can be displayed.

[0102] As an example and not limitation, a conversation timeout reminder and a transaction reminder may be set.

[0103] In the conversation timeout reminder, if the potential trading partner does not respond within the custom preset time limit when the conversation is initiated by the party, it will be considered to have timed out. In order to ensure the timeliness and accuracy of the transaction, the timed conversation will trigger a reminder and require the potential trading partner to reconfirm the bond code and transaction direction before reopening the conversation. It should be noted that the above custom set time can be set or modified according to the party's trading habits or risk control requirements.

[0104] Transaction reminder settings: After the transaction intention is reached between the two parties, when one party sends the above formatted transaction details to the trading center platform for trading, or when there is a quotation waiting to be replied by the other party on the trading center platform, relevant reminders can be set to remind the other party of the transaction progress and related to-do items. At the same time, real-time or fixed-time reminders can be set for conversations that have not yet reached a transaction in the original currency as needed.

[0105] The above-mentioned technical solution provided by the present invention automatically conducts conversations (automatic interactions) with potential transaction partners on the intention of buying and selling current securities according to the configured tasks, thereby facilitating intended transactions in the interbank current securities market. Before the conversation, relevant current securities buying and selling tasks can be formulated according to the transaction preferences of the party; during the conversation, the machine learning algorithm is used to analyze the inquiry information, identify potential transaction opportunities, and automatically generate a reply after risk verification according to the preset risk verification rules; after the conversation, the transaction elements are extracted according to the content of the conversation to assist in the final transaction, and risk verification can also be performed based on the preset risk verification rules during the process. Furthermore, the inquiry method can also initiate intelligent reminders to the dialogue party according to the time setting and the actual scenario needs during the process, and convey and display the latest developments in the transaction process, which significantly improves the inquiry efficiency of intelligent transactions.

[0106] See also Figure 3 As shown, another embodiment of the present invention also provides an inquiry device for intelligently completing spot transactions.

[0107] The price inquiry device comprises an initialization module, a dialogue module, a corpus analysis module, a risk checking module, a formatted transaction details generation module and a quotation sending module.

[0108] The initialization module is used to obtain the current securities trading tasks configured for the institution and the configured risk verification rules.

[0109] The dialogue module is used to automatically conduct a conversation with a potential transaction partner about the intention of current securities trading based on the current securities trading task configured above; the automatic conversation includes automatically initiating an inquiry to the potential transaction partner according to the current securities trading task, or after receiving the inquiry information of the potential transaction partner, automatically matching the inquiry information with the current securities trading task configured above, and automatically replying to the other party when the inquiry information can match the current securities trading task configured above.

[0110] The corpus analysis module is used to obtain the corpus content during the conversation with the potential transaction partner, determine whether the other party has a clear transaction intention based on the corpus content, and when it is determined that the other party has a clear transaction intention, confirm the transaction elements in the aforementioned conversation process based on the corpus content.

[0111] The risk checking module is used to perform risk checking on the transaction intention based on the risk checking rules configured above.

[0112] The formatted transaction details generating module is used to determine that the aforementioned transaction intention is achieved when the risk check is passed. At this time, the complete transaction element details of the transaction are collected and the formatted transaction details of the intended transaction are generated.

[0113] The quotation sending module is used to send the formal transaction to the foreign exchange trading center platform according to the formatted transaction details of the aforementioned intended transaction.

[0114] Preferably, according to the needs of risk verification, the initialization module, dialogue module, and formatted transaction details generation module can call the risk verification module to perform risk verification based on preset risk verification rules to verify whether the buying and selling tasks, inquiry responses, and transaction elements meet the risk control requirements.

[0115] In this embodiment, for the two quotation modes of dialogue quotation and request quotation, the quotation sending module is configured with different processing methods, which can be found in the description of the previous embodiment for details and will not be repeated here.

[0116] For other technical features, please refer to the description of the previous embodiment and will not be repeated here.

[0117] The present invention also provides a computer-readable storage medium for storing a computer program that can be executed by a processing unit, wherein the computer program implements the above-mentioned method when executed by the processing unit.

[0118] The storage medium may include any medium that can store program codes, such as a USB flash drive, a mobile hard disk, a read only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0119] For other technical features, please refer to the description of the previous embodiment and will not be repeated here.

[0120] In the above description, the disclosure of the present invention is not intended to limit itself to these aspects. Rather, within the scope of the target protection of the present disclosure, the components can be selectively and operatively combined in any number. In addition, terms such as "include", "encompass" and "have" should be interpreted as inclusive or open by default, rather than exclusive or closed, unless they are explicitly defined to the contrary. All technical, scientific or other terms are consistent with the meanings understood by those skilled in the art, unless they are defined to the contrary. Common terms found in dictionaries should not be interpreted too idealistically or too impractically in the context of relevant technical documents, unless the present disclosure explicitly defines them as such. Any changes and modifications made by a person of ordinary skill in the field of the present invention based on the above disclosure belong to the scope of protection of the claims.

Claims

1. A price inquiry method for intelligently completing spot transactions, characterized in that Includes steps: Obtain the current bond trading tasks configured for the institution, and the configured risk verification rules; wherein the current bond trading tasks are configured according to the transaction preferences of the institution, the transaction preferences refer to the transaction tendencies of the institution for different bonds, which are reflected by the transaction preference elements, and the current bond trading tasks configured according to the transaction preferences of the institution record one or more transaction preference elements of the institution; and the risk verification rules include a first verification rule for risk verification before conversation, when the user turns on the risk verification before automatic conversation, the verification items configured in the first verification rule are verified before adding or modifying the current bond trading tasks; the verification items include acceptable verification items and mandatory rejection verification items, for the acceptable verification items, if the current bond trading tasks do not meet the requirements of the acceptable verification items, a prompt of failure of verification is triggered, at which time, the institution can forcibly add or modify the current bond trading tasks as needed, and for the mandatory rejection verification items, if the current bond trading tasks do not meet the requirements of the mandatory rejection verification items, the current bond trading tasks cannot be added or modified; Based on the aforementioned configured current bond trading tasks that have passed the risk check, an automatic conversation is conducted with potential trading partners about the current bond trading intention, wherein the potential trading partners are selected from all the partners that conduct current bond trading transactions on the trading platform and are bond trading partners that can provide trading preference elements that match the aforementioned current bond trading tasks; the automatic conversation includes automatically initiating a price inquiry to the potential trading partner according to the aforementioned current bond trading tasks, or automatically matching the price inquiry information with the aforementioned configured current bond trading tasks after receiving the price inquiry information of the potential trading partner, and automatically replying to the other party when the price inquiry information can match the aforementioned current bond trading tasks; wherein, when automatically replying to the price inquiry of the potential trading partner, a step of calling a risk check module to perform a risk check on the replied price information is also included; Obtain the content of the corpus during the conversation with the potential transaction partner, determine whether the other party has a clear transaction intention based on the corpus content, and when it is determined that the other party has a clear transaction intention, confirm the transaction elements in the aforementioned conversation process based on the corpus content; Based on the risk check rules configured above, the risk check module performs risk check on the transaction intention; when the risk check passes, it is determined that the transaction intention is achieved. At this time, the complete transaction element details of the transaction are collected and formatted transaction details of the intended transaction are generated; According to the formatted transaction details of the aforementioned intended transaction, the formal transaction is sent to the foreign exchange trading center platform through the quotation sending module.

2. The method according to claim 1, characterized in that: The configured current bond trading tasks are single tasks or batch tasks; The transaction preference elements include transaction direction, transaction type, bond code, settlement speed, target volume, single transaction volume, effective time, price, whether to send price and / or whether to reply to the other party's inquiry.

3. The method according to claim 1, characterized in that When confirming the transaction elements in the aforementioned conversation process based on the corpus content, calling a preset transaction data recognition model to perform natural language analysis on the corpus content to identify the transaction element information in the corpus; The transaction data recognition model adopts a combination of the BERT-CRF model and transaction rules when identifying transaction elements. The transaction rules include the transaction rules of the interbank bond market. The type, quantity and / or format of the transaction elements extracted by the transaction data recognition model meet the requirements of the transaction rules.

4. The method according to claim 1, characterized in that: The steps of judging whether the other party has a clear transaction intention and confirming the transaction elements based on the content of the corpus include: When this party initiates an inquiry, the transaction elements of the inquiry sent to the potential transaction partner include at least the bond code, transaction volume and settlement speed; at this time, the corpus content of the conversation between this party and the potential transaction partner is identified to determine whether the corpus contains content that the potential transaction partner expects to conclude a transaction and / or quotation content with clear transaction elements. If the above content is included, it is determined that the potential transaction partner has a clear transaction intention; at this time, the steps of confirming the transaction elements are as follows: this party sends a request to the potential transaction partner to confirm the relevant transaction elements, and after receiving the potential transaction partner's confirmation of the above request, the above transaction elements are confirmed; When we receive an inquiry from a potential transaction partner, we determine whether the bond code and transaction direction of the inquiry match the aforementioned current bond buying and selling tasks, and if they match, we determine that the potential transaction partner has a clear transaction intention; at this time, the steps to confirm the transaction elements are as follows: we send a request to the potential transaction partner to supplement other relevant transaction elements, and after receiving the supplementary transaction element information from the potential transaction partner, we integrate the supplemented transaction elements with the existing transaction elements and then confirm the transaction elements.

5. The method according to claim 1, characterized in that Corresponding to the method of receiving the inquiry information of the potential transaction object, the steps of automatically replying to the other party include: Collect the identity information of the potential transaction object and determine whether the potential transaction object is on the blacklist preset by the institution. If it is determined that the potential transaction object is on the blacklist, the automatic conversation process ends; otherwise, proceed to the next step; After collecting the bond code and transaction direction inquired by the potential transaction object, match them with the aforementioned current bond trading task. If it is determined that it cannot be matched with the aforementioned current bond trading task, it will prompt the potential transaction object that there is no quotation from this party; otherwise, proceed to the next step; Collect the price and transaction volume information provided by the potential transaction partner when inquiring, and determine whether the provided price and transaction volume information is unfavorable to the party. If it is determined that the provided price or transaction volume is unfavorable to the party, prompt the potential transaction partner that the transaction cannot be concluded; otherwise, proceed to the next step; We will automatically reply to the above inquiry, and the automatic reply information will include our quotation information.

6. The method according to claim 1, characterized in that: When the risk check module is used to check the risk of the aforementioned transaction intention, the risk check is performed according to the quotation strategy configured by the party; At this time, a price strategy setting module is configured to configure the quotation strategy of the party; the market price benchmark in the quotation strategy selects the quotation quotation, and accepts the BP deviation setting of the Bid Spread / Ofr Spread based on the quotation quotation. After setting, the recommended price is automatically obtained when quoting; According to the quotation strategy, the risk check module is configured as follows: when the other party's quoted price is better than the price set by our party, the check is judged to be passed, our party accepts the quotation and proceeds to the next process; when the other party's required quantity is greater than the target quantity or the quoted price is unfavorable to our party, the check is judged to be failed, a reminder that the transaction intention cannot be reached is issued, and when the preset conditions are met, the reason why the transaction intention cannot be reached is sent to the other party.

7. The method according to claim 1, characterized in that: When performing risk check on the replied price information; the check items configured in the risk check module include price deviation check item, price deviation value check item, quotation quantity check item and transaction volume check item; At this time, the risk verification module is configured to: collect the quotation information in the reply information of the party, determine whether the aforementioned quotation information meets the limit range of the aforementioned verification items, and when the quotation information meets the limit range of the aforementioned verification items, determine that the verification is passed, and send the reply information of the party to the aforementioned potential transaction object; when the quotation information does not meet the limit range of one or more of the aforementioned verification items, determine that the verification fails, prompt the potential transaction object that the party has no quotation yet, and terminate the transaction process.

8. The method according to claim 1, characterized in that: The formatted transaction details are a formatted transaction intention form to be executed. The formatted transaction intention form is structured data in a preset data format and includes necessary transaction elements required under the inter-bank bond market transaction rules.

9. The method according to claim 8, characterized in that: Including a dialogue quotation mode and a request quotation mode, according to different quotation modes, the quotation sending module is configured as follows: In the dialogue quotation mode, when the formatted transaction intention form is initiated by this party, the aforementioned formatted transaction details are collected and the quotation for the formal transaction is sent to the foreign exchange trading center platform; when the formatted transaction intention form is initiated by the potential trading counterparty, after the party identifies the source of the received quotation information, it confirms the formal transaction to the foreign exchange trading center platform; In the quotation request mode, when we are the party requesting the quotation, we initiate the sending of the formatted transaction intention form, collect the aforementioned formatted transaction details, and send the formal transaction to the foreign exchange trading center platform; when we are the party requesting the quotation, the formatted transaction intention form can be converted into a proposed quotation order. At this time, after identifying the quotation information received by us, we match the proposed quotation order with the front-end transaction according to the elements. After the match is successful, we reply to the relevant elements to the inquirer. After the inquirer agrees to the transaction, we confirm the formal transaction with the foreign exchange trading center.

10. The method according to claim 1, characterized in that: Throughout the transaction process, reminders are sent to potential transaction partners based on preset time limits and / or reminder scenarios, and the latest dynamic information of the transaction process is displayed.

11. A price inquiry device for intelligently completing spot bond transactions according to the method of claim 1, characterized in that include: Initialization module, used to obtain the current securities trading tasks configured for the institution, as well as the configured risk verification rules; A dialogue module is used to automatically conduct a conversation with a potential trading partner about the intention of trading a current security based on the aforementioned configured current security task; the automatic conversation includes automatically initiating a price inquiry to the potential trading partner according to the aforementioned current security task, or automatically matching the price inquiry information with the aforementioned configured current security task after receiving the price inquiry information of the potential trading partner, and automatically replying to the other party when the price inquiry information can match the aforementioned current security task; wherein, when automatically replying to the price inquiry of the potential trading partner, it also includes the step of calling the risk check module to perform a risk check on the price information of the reply; a corpus analysis module is used to obtain the corpus content during the conversation with the potential trading partner, determine whether the other party has a clear trading intention based on the corpus content, and when it is determined that the other party has a clear trading intention, confirm the transaction elements in the aforementioned conversation based on the corpus content; A risk check module, used to perform risk check on the aforementioned transaction intention based on the aforementioned configured risk check rules; A formatted transaction details generation module is used to determine that the aforementioned transaction intention has been reached when the risk check has passed. At this time, the complete transaction element details of the transaction are collected and the formatted transaction details of the intended transaction are generated; The quotation sending module is used to send the formal transaction to the foreign exchange trading center platform based on the formatted transaction details of the aforementioned intended transaction.

Images