Patent Application
20170286912
The present invention relates to the field of interaction between electronic devices and human users, more particularly, to the field of electronic device-based intelligent reminding, and still more specifically, to intelligent reminder methods and systems.
In recent years, with the widespread adoption of smart phones, it has become an increasingly common practice of merchants to disseminate their information via Short Message Service (SMS) or social media accounts. In these push messages, in addition to those indeed containing information of users' interest or serving as reminders to users such as, for example, delivery notification messages, railway ticket booking confirmation messages, messages serving as proofs of purchase of group purchase coupons and phone bill messages, there are also a huge number of spam messages. As a result, users' cell phones or electronic communication terminals are often flooded by numerous advertising or other spam messages which are not reminder messages, and the users usually have to spend a lot of time to read a great number of such specious push messages in order to identify those really useful to them. This does not only lead to a reduced efficiency but also often causes omission of important messages and thus inconvenience in the users' work or lives.
In addition, the merchants may carry out the notification or reminding using different software. For instances, some of the merchants uses SMS messages, and there are also some relying on Facebook or Wechat accounts or emails. For the users, there is not any versatile mechanism for handling all the push messages in diversified forms delivered from distinct merchants, not to mention a versatile reminding mechanism established based on the push messages in various forms.
In order to overcome the above-described problems, there is a need for a more intelligent reminder method, system and apparatus for use in electronic devices.
The technical problem to be solved by the present invention is to obtain reminder messages from push messages received by an electronic device and realize intelligent reminding.
According to one aspect of the invention, there is provided an intelligent reminder method, comprising: obtaining push messages received by an electronic device terminal, wherein the push messages comprises reminder messages, the reminder messages comprising classification-related feature data; based on the classification-related feature data, filtering the push messages to obtain the reminder messages and labelling the reminder messages into different categories, wherein each of the reminder message categories is associated with at least one action; and based on the different reminder message categories, carrying out the actions associated with the reminder message categories.
According to another aspect of the invention, there is further provided an intelligent reminder system comprising: message reception means, for receiving push messages associated with predetermined accounts; detection means, for, based on classification-related feature data in the push messages, examining the push messages to obtain reminder messages and labelling the reminder messages into different categories, wherein reminder messages in each of the categories are associated with at least one action; and reminder means for reminding a user based on the reminder messages.
Compared to the prior art, the present invention fully takes into account the characteristics of reminder messages, subjects the push messages to processing based on classification-related feature data so as to obtain reminder messages therefrom, classifies the reminder messages into different categories, and carries out actions corresponding to the different categories according to user inputs, thereby achieving intelligent reminding based upon the push messages.
Other features, objects and advantages of the invention will become more apparent upon reading the detailed description of several non-limiting embodiments taken in conjunction with the accompanying drawings, in which:
Exemplary embodiments of intelligent reminder methods and systems according to the present invention will be explained more fully below with reference to the accompanying drawings in which the same reference symbols in different drawings are used to indicate similar or identical items. Although a few exemplary embodiments and features of the present invention are set forth below, modifications, alterations and other substitutions made to the invention without departing the concept thereof such as, for example, equivalent substitutions, additions or modifications to element(s) illustrated in the drawings, or substitutions, rearrangements or additions of steps, shall not be construed as limiting the present invention, and the proper scope of the invention shall be as defined in the appended claims.
According to some embodiments, intelligent reminder systems according to the present invention can be configured to examine push messages received in an electronic device through extracting classification-related feature data from the push messages, identify those of the push messages meeting criteria as reminder messages for notifying a user, group the reminder messages into different categories based on the classification-related feature data, and in the event of the user selecting one of the reminder messages, based on the category of the selected reminder message, carry out action(s) associated with its category.
The electronic device 110 may be either connected to a telecommunication network utilizing, for example, CDMA, 2G 3G or 4G schemes and receive push messages 101 from the telecommunication network, or connected to the Internet through a broadband connection such as, for example, a ADSL, VDSL, fiber optic, wireless, cable TV or satellite connection, or through a narrowband connection such as, for example, a PSTN, GPRS, 2G or 3G connection and receive push messages 101 associated with some predefined accounts.
The push messages 101 referenced herein may be Short Message Service (SMS) messages such as, for example, advertising or reminder messages sent from merchants via base stations to a Short Message Service Center (SMSC) and further to the user's cell phone via a GMS network or an SMS gateway, so that the user can receive and open the messages with the cell phone and view them on a screen thereof. The push messages 101 may also be messages associated with predefined accounts such as, for example, those sent to a certain e-mail address or to an iTune, Facebook, Wechat, QQ, or other account. After the messages reach a designated electronic device or account, the user can receive them by the device or account.
Subsequently, detection means 120 or cloud detection means 130 may examine the push messages 101 received in the electronic device 110 by extracting classification-related feature data therefrom such as to preclude non-reminder messages and obtain reminder messages and label the reminder messages into different categories, and reminder means 140 may notify the user, wherein reminder messages in each of the categories are associated with at least one actions.
According to some embodiments of the present invention, the detection means 120 may parse all SMS messages obtained in the cell phone, or a predetermined proportion such as, for example, from 1% to 30%, of the messages, or those of the messages meeting predetermined filtration criteria such as, for example, only those from non-phonebook contacts, or only those containing predetermined contents such as, for example, “balance”, “Yuan” or “remaining balance”, or only those from predetermined sender numbers such as, for example, 100861.
According to some embodiments, from the SMS messages to be parsed, the detection means 120 can extract classification-related feature data which can identify of different message categories, such that reminder messages can be obtained and labelled into different categories.
In one embodiment, the detection means 120 may label the SMS messages to be parsed into different categories depending on whether they contain predetermined classification-related feature data or not. The classification-related feature data may include strings consisting of keywords in the messages or synonymous words or phrases thereof, or be message sender numbers, Short Message Base Station Service Center codes, or a combination of these feature data.
For example, the detection means 120 can label SMS messages as payment reminder messages when detecting therein the presence of, for example, strings containing the keywords “Yuan”, “remaining balance”, “less than” or synonymous words or phrases thereof, or as delivery notification messages when detecting therein the presence of, for example, strings containing the keywords “express delivery”, “includes”, “shipped out”, “tracking number” or synonymous words or phrases thereof.
As another example, upon the detection means 120 extracting the sender number from a SMS message as, for example, “1-800-604-9961”, it may identify the number as the customer service number of the Bank of America and accordingly label the message as a bank notification message.
As a further example, Short Message Base Station Service Center codes may also serve as the classification-related feature data, according to which, the detection means 120 may examine the push messages so as to obtain the reminder messages and then label them.
In another embodiment, the detection means 120 may also determine proportions of the classification-related feature data corresponding to the different categories in the SMS messages to be parsed and thereby accomplish labelling the messages into the categories.
In still another embodiment, the detection means 120 may also perform semantic analysis or regular expression matching on the messages to determine whether there are contents identical or similar to classification-related feature data corresponding to predetermined categories and, if positive, label the messages into the respective categories.
The detection means 120 can further detect an input signal of the user and, in the event of the signal indicative of the user's selection of one of the reminder messages, carry out action(s) 150 associated with the category to which the selected reminder message belongs. For example, with the reminder means 140 having reminded the user and in the case of the user being detected to select a payment reminder message, action(s) associated with the category “Payment” may be obtained and carried out. For example, a link to a corresponding payment interface may be opened to allow the user to directly fulfill the payment or recharge requirements.
According to some embodiments, the detection means 120 or cloud detection means 130 may be an executable program that can be read and run by a computer processor. For example, the detection means 120 may be deployed in the electronic device 110 and receive and obtain the push messages 101 through monitoring incoming messages of the electronic device 110. As another example, the cloud detection means 130 may be deployed in a cloud server and obtain the push messages 101 received in the electronic device 110 from network transmissions. According to some other embodiments, the intelligent reminder system 100 may include both of the local detection means 120 and the cloud detection means 130 which are configured to successively or simultaneously examine the push messages received in the electronic device 110, depending on system settings or network conditions.
The reminder messages are further transmitted to the reminder means 140 which then reminds the user based on the messages in the form of videos or audios. For example, the reminder means 140 may comprise one or more textual or graphical screens or other display devices and programs for driving the display devices and display the reminder messages to the user, or comprise a sound device such as a speaker or a program for driving the sound device and present the reminder messages in the form of speech, or comprise means for performing other reminding functions, for example, by indication lamps, vibration, etc.
The processor 220 may be a Central Processing Unit (CPU) or a Graphics Processing Unit (GPU). Specifically, the processor 220 may further comprise one or more printed circuit boards (PCBs) or microprocessor chips configured to perform computer program instruction sequences so as to implement a variety of methods described in further detail below. In some embodiments, the processor 220 may be configured to receive push messages from the reception means 210, filter the push messages based on classification-related feature data contained therein to obtain reminder messages from the push messages, store the reminder messages in the memory 240, label the reminder messages into different categories and monitor whether any of the reminder messages is selected. In addition, upon selection of one of the reminder messages being detected, action(s) associated with the category of the selected reminder message is obtained from the memory 240 and carried out.
The memory 240 may comprise one or more of a random access memory (RAM) and a read-only memory (ROM). Computer program instructions to be executed by the processor 220 may be accessed or read from the ROM or any other suitable memory location and loaded in the RAM. For example, the memory 240 may store one or more software applications. The software applications stored in the memory 240 may include operating systems for conventional computer systems and software-controlled devices. In addition, the memory 240 may store an entire software application or only part thereof executable by the processor 220. For example, the memory 240 can store intelligent reminder software executable by the processor 220 and implement the intelligent reminder method.
In some embodiments, the memory 240 may also store one or more of master data, user data, application data and program codes. For example, the memory 240 can store a local database 330. In some embodiments, the local database 330 may include one or more reminder messages. For example,
In some embodiments, the reception means 210 and the reminder means 140 may be coupled to the processor 220 by suitable interface circuitry. In some embodiments, the reception means 210 may be means for receiving SMS messages or mails or reception means for use with other communications software.
According to some embodiments, the intelligent reminder system 100 can further comprises a communication interface 231 which can provide a communication connection allowing information exchanges between the intelligent reminder system 100 and some external devices. According to one embodiment, the communication interface 231 may include a network interface (not shown) configured to transmit information to a cloud service 230 and receive information therefrom. According to some embodiments, the cloud service 230 may be implemented as a web service on the Internet, a cloud storage service or the like.
Referring to
In one specific embodiment, the reception means 210 receives N push messages within a first time threshold and transmits these push messages to the filtration means 310.
After that, the filtration means 310 filters all of the push messages in a predetermined manner so as to obtain reminder messages therefrom.
The filtration means 310 may filter the push messages based on classification-related feature data. For example, the filtration means 310 may filter the push messages based on SMS message sender numbers such as to preclude those of the push messages from phonebook contacts and identify those from non-phonebook contacts as reminder messages. Alternatively, the filtration means 310 may also filter the push messages based on keywords in SMS messages or their combinations. For example, the filtration means 310 may identify those of the SMS messages whose text contains the predetermined keywords “balance”, “amount” and “insufficient” or synonymous words or phrases thereof as the reminder messages. Still alternatively, the filtration means 310 may also filter the push messages based on Short Message Base Station Service Center codes. For example, the filtration means 310 may preclude some push messages from pseudo base stations, in order to avoid reminding the user with fraud SMS messages as the reminder messages. In addition, the filtration means 310 may also perform repeated filtration based on classification-related feature data of multiple types.
In another embodiment, filtration means 310 may also filter a predetermined proportion of the push messages. For example, the filtration means 310 may perform the classification-related feature data-based filtration on 1% to 30% of the push messages.
The reminder messages filtered from the filtration means 310 are further labelled into different categories by the labelling means 320 also based on the classification-related feature data.
The filtration means 310, labelling means 320 and database 330 may be deployed locally, for example, on the same user terminal as the reception means 210 and reminder means 140. According to some other embodiments, referring to
According to some embodiments, referring to
In one embodiment, the classification means 410 may further comprise a first classifier which filters and examines the push messages and in the event of having detected the presence of predetermined classification-related feature data in a push message, labels the push message as a reminder message belonging to a category associated with the classification-related feature data.
For example, referring to
According to some embodiments, the first classifier may consecutively filter the push messages based on their categories. For example, the first classifier may first determine whether a received SMS message is a reminder message belonging to the category “ISP” and, if negative, continue to determine whether it is a reminder message belonging to other categories, until one of the categories corresponding to the SMS message is obtained. In one embodiment, when the push message is found to not belong to any of the categories, it is determined as a non-reminder message.
According to some embodiments, the first classifier can further comprise an extraction module and a determination module. In one embodiment, the extraction module may extract the classification-related feature data from the push messages according to types of the data, followed by the determination module performing the determination in a consecutive manner. For example, the extraction module may first extract sender numbers from all SMS messages, followed by the determination module determining whether the sender numbers of the SMS messages belong to the category “ISP”, i.e., containing “10086?”, or “10001?”, or “10011?”. After that, the extraction module may extract Short Message Base Station Service Centercodes or message text from those of the SMS messages with their sender numbers not belonging to the category “ISP”, followed by the determination module performing the corresponding determination, until the categories of the SMS messages have been obtained. In another embodiment, the first classifier may perform the determination in a manner of one SMS message after another. For example, the extraction module may extract sender numbers, Short Message Base Station Service Center codes and message text corresponding to the SMS messages, followed by the determination module determining the categories to which the SMS messages belong based on the extracted data. In some embodiments, the determination module may perform direct determination based on a certain type of classification-related feature data or a certain classification-related feature datum. For example, when the determination module has detected the keyword “railway” or “flight” in the text of a SMS message, it can determine the SMS message as a reminder message of the category “Ticketing”. In some embodiments, the determination module may perform the determination based on a combination of several classification-related feature data. For example, the determination module may determine a SMS message as a reminder message of the category “Ticketing” only when its text contains both “balance” and “less than” and its sender number is “10086”.
According to some embodiments, the classification means 420 may further comprise a second classifier which extracts classification-related feature data from received push messages, calculates probabilities of the push messages belonging to the categories and labels the push messages into the respective corresponding categories.
For example, in the case of the second classifier extracting classification-related feature data from an i-th push message as {x1i, x2i, . . . , xni} as the category A, category B, . . . , and category n correspond to weight values A, weight values B, . . . , and weight values n, respectively, where the weight values A are {wA0, wAi, . . . , wAn}, the weight values B are {wB0, wB1, . . . , wBn}, . . . , and the weight values n are {wn0, wn1, . . . , wnn}, the second classifier can calculate probabilities of the push message belonging to the respective categories.
Specifically, the second classifier may calculate a probability of the push message belonging to the category A as:
where, fA(xi)=wA0+wA1·x1i+wA2·x2i+ . . . +wAn·xni.
Similarly, a probability of the push message belonging to the category B, calculated by the second classifier, may be:
where, fB(xi)=WB0+WB1·x1i+wB2x2i+ . . . +wBn·Xni.
At last, a probability of the push message belonging to the category n, calculated by the second classifier, may be:
where, fn(xi)=wn0+wn1·x1i+wn2·x2i+ . . . +wnn·xni.
Afterward, based on these probabilities of the push message belonging to the respective categories, the second classifier may label the push message into the one of the categories having the greatest probability. In other words, when PA=max{PA, PB, . . . , Pn}, the push message belongs to the category A.
In another embodiment, the classification means 410 may comprise a semantic analyzer configured to perform context-related examination on the push messages in terms of their structures so as to determine their categories. In yet another embodiment, the classification means 410 may comprise a matching module configured to perform regular expression matching on the push messages to determine whether they meet filtering logics of the regular expressions.
According to some embodiments, referring to
In another embodiment, the updating means 340 may also update the actions corresponding to the reminder messages.
After being obtained, the reminder messages and their corresponding categories are transmitted to the reminder means 140.
According to some embodiments, before the reminder messages and categories are sent to the reminder means 140, each of the reminder messages may be formed, by the structuring means 420, into a predetermined structure comprising the corresponding category.
For example, after being processed by the classification means 410, a push message from the number “10086”, saying “The remaining account balance for your phone number 138xxxxxxxx is 5.76 Yuan, please recharge the account in a timely way to prevent undesirable disconnection of your phone. Thank you for your cooperation”, may be obtained as a reminder message belonging to the category “ISP” and the structuring means 420 may form the reminder message, based on its content, into a predetermined structure such as, for example, a JSON formatted data schema:
Structuring the reminder messages with the structuring means 420 allows other reception programs not designed for the reminder messages to receive and process such data, thereby generating uniform reminder indicators.
In one embodiment, in case of the reminder means 140 accomplishing the reminding in a manner of display, the structuring means 420 may structure the reminder messages depending on display means used by the reminder means 140 such as, for example, a display program for the reminder messages, such that the reminder messages can be read by the display program and displayed by a display device 140.
Subsequently, after receiving the structured reminder messages, the reminder means 140 can provide the reminder messages with different reminder indicators according to the respective categories to which they belong, so as to remind the user. The reminder indicators may be names, descriptions or so forth of the categories corresponding to the reminder messages.
The reminder means 140 may accomplish the reminding by using either an interface of an existing application program such as, for example, Messaging or Notebook, or an otherwise designed interface. Referring to
Referring to
According to some embodiments, the monitoring and execution means 520 may detect a further input of the user, acquire actions corresponding to reminder messages selected thereby and carry out the actions. In some embodiments, the monitoring and execution means 520 may include input monitoring means 521, action obtention means 522 and action execution means 523. The input monitoring means 521 may detect whether there is an input of the user. For example, when it detects that the user has performed a tapping, checking or another inputting operation that will result in the selection of one of the reminder messages on the screen by means of a finger or an input device such as a stylus, it may record a user input signal and transmit the recorded user input signal to the action obtention means 522. The action obtention means 522 may then determine, based on the received user input signal such as, for example, a coordinate or a pixel touched by the user, the reminder message selected by the user and obtain action(s) corresponding to that reminder message. The action execution means 523 may then carry out the action(s) obtained by the action obtention means 522, such as for example, opening an associated link, picture or text.
According to some embodiments, referring to
According to some embodiments, referring to
According to some embodiments, the action execution means 523 may perform further selection based on the user's input. Referring to
Referring to
According to some embodiments, in an Android system, the push messages may be obtained by monitoring SMS messages of the system.
With the push messages having been obtained, based on the classification-related feature data, in one embodiment, the push messages may be filtered to obtain the reminder messages, and the reminder messages obtained from the filtration may be labelled, according to the classification-related feature data, into different categories. For example, in case of SMS messages, SMS messages can be labelled into different categories according to their sender numbers, contents, Short Message Base Station Service Center codes, and so forth. In another embodiment, the push messages may be classified to the categories based on the classification-related feature data, with those not belonging to any of the categories labelled as non-reminder messages.
Specifically, in one embodiment, the push messages may be examined, with those thereof containing the classification-related feature data that indicate predetermined categories labeled as reminder messages of the respective corresponding predetermined categories. In another embodiment, in case of the classification-related feature data being hybrid data, probabilities of the push messages belonging to the different categories may be calculated such that the push messages are labelled as reminder messages of the respective most probable categories. In yet another embodiment, semantic analysis may be performed on the contents of the push messages such as to label the push messages as reminder messages of the predetermined categories. In still another embodiment, regular expression matching may be performed on the contents of the push messages to determine whether there are contents identical or similar to the classification-related feature data of the predetermined categories.
According to some embodiments, the step S2 may further include an updating step for updating correspondence relationships between the classification-related feature data and the labelling categories or for updating the actions associated with the labelling categories.
With the reminder messages and their associated categories having been obtained, the actions associated with their categories can be carried out. Specifically, after the reminder messages have been presented to the user, an input of the user may be further detected, based on which a selected one of the reminder messages can be determined. Action(s) and action data associated with the selected reminder message may then be obtained, followed by execution of the associated action(s).
In one embodiment, the method may further comprise structuring the reminder messages and notifying the user in a predefined order and manner based on the different reminder message categories, after the obtention of the reminder messages.
Compared to the prior art, the present invention fully takes into account the characteristics of reminder messages, subjects the push messages to processing based on classification-related feature data so as to obtain reminder messages therefrom, classifies the reminder messages into different categories, and carries out actions corresponding to the different categories according to user inputs, thereby achieving intelligent reminding based upon the push messages.
While specific embodiments have been described above, it is to be understood that the present invention is not limited to the disclosed embodiments. Those skilled in the art can make various variations or modifications within the scope defined by the appended claims, which are, however, deemed not to affect the essence of the present invention.
| Number | Date | Country | Kind |
|---|---|---|---|
| 201410477689.7 | Sep 2014 | CN | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/CN2015/089909 | 9/18/2015 | WO | 00 |
