CN107341233B

CN107341233B - Position recommendation method and computing device

Info

Publication number: CN107341233B
Application number: CN201710534021.5A
Authority: CN
Inventors: 张朝阳; 谢双宾; 郭旺平
Original assignee: Beijing Lagou Technology Co ltd
Current assignee: Beijing Lagou Technology Co ltd
Priority date: 2017-07-03
Filing date: 2017-07-03
Publication date: 2020-11-06
Anticipated expiration: 2037-07-03
Also published as: CN107341233A

Abstract

The invention discloses a position recommendation method, which is used for recommending positions related to target positions and is executed in computing equipment, and comprises the following steps: calculating the co-occurrence times of the target position and other positions according to the behavior logs of the users; sequentially selecting a first number of positions according to the sequence of the common times from high to low to generate a recommendation candidate set of target positions; respectively calculating the text similarity of the target position and each position in the recommendation candidate set; and selecting the positions with the text similarity value larger than the threshold value for recommendation. The invention also discloses corresponding computing equipment.

Description

Position recommendation method and computing device

Technical Field

The invention relates to the field of data processing, in particular to a position recommendation method and computing equipment.

Background

With the development of internet technology, network application/recruitment becomes a main approach for job seekers to apply work and recruit employees by employing units. The personnel units and job seekers register accounts on the talent recruitment website of the third party and search and communicate with each other to search for the most satisfactory objects.

The talent recruitment website has a large number of positions, and after an applicant logs in the website, the positions can be browsed from the position list, and the positions which can be suitable for the applicant can be searched by means of keywords. Generally, the number of job listings is huge, and the applicant cannot find the job in which the applicant is interested in a short time; and the keyword search is used, the number and the accuracy of the keywords used by the applicant are limited, the number of searched positions is often very large, and the position suitable for the applicant is difficult to find quickly. In order to solve the problem, the method commonly adopted by the recruitment website at present is as follows: and recommending positions for the applicants, and actively pushing the positions possibly suitable to the applicants by the website.

The current position recommendation method mainly adopts a method based on statistics, namely, the browsing times and the concerned degree of all positions are counted to be used as the popularity of the positions, and then the position with the highest popularity is recommended to an applicant. However, resume deliveries received by a plurality of positions on the recruitment website are few, or the resume deliveries are still 0 for a position which is just released for a long time, and in this case, the position recommendation method has the problems of inaccurate recommendation and low coverage rate.

Therefore, a position recommendation method capable of solving the problem of inaccurate recommendation due to data sparsity is required.

Disclosure of Invention

To this end, the present invention provides a position recommendation method and computing device in an attempt to solve or at least alleviate at least one of the problems identified above.

According to one aspect of the present invention, there is provided a position recommendation method for recommending a position associated with a target position, the method being performed in a computing device, comprising the steps of: calculating the co-occurrence times of the target position and other positions according to the behavior logs of the users; sequentially selecting a first number of positions according to the sequence of the common times from high to low to generate a recommendation candidate set of target positions; respectively calculating the text similarity of the target position and each position in the recommendation candidate set; and selecting the positions with the text similarity value larger than the threshold value for recommendation.

Optionally, in the recommendation method according to the present invention, the step of calculating the number of co-occurrences of the target position and the other positions according to the behavior logs of the plurality of users includes: acquiring post delivery records of a plurality of users, and respectively calculating the delivery co-occurrence times of a target post and each of other posts according to the post delivery records; acquiring the position browsing records of a plurality of users, and respectively calculating the browsing co-occurrence times of the target position and each of other positions according to the position browsing records; and respectively calculating the co-occurrence times of the target position and each of other positions according to the delivery co-occurrence times and the browsing co-occurrence times.

Optionally, in the recommendation method according to the present invention, the step of calculating the number of co-occurrences between the target position and each of the other positions according to the number of delivery co-occurrences and the number of browsing co-occurrences includes: multiplying the delivering co-occurrence times and the browsing co-occurrence times by a first coefficient and a second coefficient respectively to obtain the treated delivering co-occurrence times and the processed browsing co-occurrence times; and adding the treated delivery co-occurrence times and the treated browsing co-occurrence times to obtain the co-occurrence times.

Optionally, in the recommendation method according to the present invention, the step of calculating the text similarity between the target position and each position in the recommendation candidate set respectively includes: calculating a feature vector of each position, wherein the feature vector is used for describing the probability distribution of the position on each topic; and calculating the cosine similarity of the included angle between the feature vector of the target position and the feature vector of each position in the recommended candidate set as the text similarity.

Optionally, in the recommendation method according to the present invention, the step of calculating the feature vector of each position includes: and generating a feature vector of each position by using the document theme generation model.

Optionally, in the recommendation method according to the present invention, the method further includes the step of training to generate a document theme generation model: randomly assigning a theme to each word in each position in the training sample as an initial theme; scanning a training sample, and sampling by using a Gibbs sampling method to obtain the probability that each word belongs to each theme; and repeating the sampling step until the probability tends to be stable, so as to obtain the document theme generation model.

Optionally, in the recommendation method according to the present invention, the step of training the document theme generation model further includes: and determining a theme set of the training sample according to the theme covered by each position in the training sample.

Alternatively, in the recommended method according to the present invention, the first coefficient is 1, and the second coefficient is 0.2.

According to another aspect of the present invention, there is provided a computing device comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the methods described above.

According to a further aspect of the invention there is provided a computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform any of the methods described above.

According to the job recommendation scheme, the relevance between the target job and other candidate jobs is described by using the co-occurrence times, specifically, the problem of low coverage rate caused by too little delivery data is considered, browsing data is added to expand data volume, the co-occurrence times are calculated according to the delivery co-occurrence times and the browsing co-occurrence times between jobs, and the problem of data sparseness is solved to a great extent.

In addition, text similarity between positions is calculated by using an LDA model, and a recommended candidate set obtained by co-occurrence times is filtered once in the aspect of text similarity, so that the recommendation accuracy is further improved.

Drawings

To the accomplishment of the foregoing and related ends, certain illustrative aspects are described herein in connection with the following description and the annexed drawings, which are indicative of various ways in which the principles disclosed herein may be practiced, and all aspects and equivalents thereof are intended to be within the scope of the claimed subject matter. The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description read in conjunction with the accompanying drawings. Throughout this disclosure, like reference numerals generally refer to like parts or elements.

FIG. 1 shows a schematic diagram of a computing device 100, according to an embodiment of the invention; and

fig. 2 shows a flow diagram of a job recommendation method 200 according to one embodiment of the invention.

Detailed Description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

Fig. 1 is a block diagram of an example computing device 100. In a basic configuration 102, computing device 100 typically includes system memory 106 and one or more processors 104. A memory bus 108 may be used for communication between the processor 104 and the system memory 106.

Depending on the desired configuration, the processor 104 may be any type of processing, including but not limited to: a microprocessor (μ P), a microcontroller (μ C), a Digital Signal Processor (DSP), or any combination thereof. The processor 104 may include one or more levels of cache, such as a level one cache 110 and a level two cache 112, a processor core 114, and registers 116. The example processor core 114 may include an Arithmetic Logic Unit (ALU), a Floating Point Unit (FPU), a digital signal processing core (DSP core), or any combination thereof. The example memory controller 118 may be used with the processor 104, or in some implementations the memory controller 118 may be an internal part of the processor 104.

Depending on the desired configuration, system memory 106 may be any type of memory, including but not limited to: volatile memory (such as RAM), non-volatile memory (such as ROM, flash memory, etc.), or any combination thereof. System memory 106 may include an operating system 120, one or more applications 122, and program data 124. In some embodiments, application 122 may be arranged to operate with program data 124 on an operating system. The program data 124 includes instructions, and in the computing device 100 according to the present invention, the program data 124 contains instructions for performing a job recommendation method.

Computing device 100 may also include an interface bus 140 that facilitates communication from various interface devices (e.g., output devices 142, peripheral interfaces 144, and communication devices 146) to the basic configuration 102 via the bus/interface controller 130. The example output device 142 includes a graphics processing unit 148 and an audio processing unit 150. They may be configured to facilitate communication with various external devices, such as a display or speakers, via one or more a/V ports 152. Example peripheral interfaces 144 may include a serial interface controller 154 and a parallel interface controller 156, which may be configured to facilitate communication with external devices such as input devices (e.g., keyboard, mouse, pen, voice input device, touch input device) or other peripherals (e.g., printer, scanner, etc.) via one or more I/O ports 158. An example communication device 146 may include a network controller 160, which may be arranged to facilitate communications with one or more other computing devices 162 over a network communication link via one or more communication ports 164.

A network communication link may be one example of a communication medium. Communication media may typically be embodied by computer readable instructions, data structures, program modules, and may include any information delivery media, such as carrier waves or other transport mechanisms, in a modulated data signal. A "modulated data signal" may be a signal that has one or more of its data set or its changes made in such a manner as to encode information in the signal. By way of non-limiting example, communication media may include wired media such as a wired network or private-wired network, and various wireless media such as acoustic, Radio Frequency (RF), microwave, Infrared (IR), or other wireless media. The term computer readable media as used herein may include both storage media and communication media.

Computing device 100 may be implemented as a server, such as a file server, a database server, an application server, a WEB server, etc., or as part of a small-form factor portable (or mobile) electronic device, such as a cellular telephone, a Personal Digital Assistant (PDA), a personal media player device, a wireless WEB-watch device, a personal headset device, an application specific device, or a hybrid device that include any of the above functions. Computing device 100 may also be implemented as a personal computer including both desktop and notebook computer configurations. In some embodiments, the computing device 100 may also be implemented as a server configured to perform the position recommendation method 200 according to the present invention.

Generally, when a user browses certain position information on a recruitment website, some areas are reserved on a page of the position information for recommending one or more related positions for the user, for example, when the user browses position a, positions B, C, D related to position a and interested by the user can be recommended for the user according to characteristics of position a. At this time, we refer to position a as the target position, and position B, C, D, etc. are the recommended positions associated with target position a.

Fig. 2 shows a flow diagram of a job recommendation method 200 according to one embodiment of the invention. The position recommendation method 200 is suitable for execution in a computing device implemented as a server, such as the computing device 100 shown in fig. 1.

The method begins in step S210, and calculates the number of co-occurrences between the target position and other positions according to the behavior logs of the plurality of users. The "co-occurrence count" is, as the name implies, the number of times the target position co-occurs with other positions. Specifically, step S210 may be divided into the following three steps:

first, historical delivery data is collected. When a position is delivered by a user for a plurality of times, resumes or shares the position with other positions for a plurality of times, the position is considered to have high correlation with other positions and high recommended probability, so that position delivery records of a plurality of users are obtained, and the delivery co-occurrence times of the target position and each other position are respectively calculated according to the obtained position delivery records. For example, the post delivery records obtained for a plurality of users are shown in table 1:

TABLE 1 post delivery record sheet

	P1	P2	P3	P4	P5
						U1		1	1
U2	1	1	1		1
						U3	1	1	1
U4	1			1	1

Where U represents a user and P represents a post, then each row in table 1 represents post delivery records for each user, each column represents a record for each post delivered, and the numerical value represents the number of deliveries, e.g., the number "1" in the second row and third column represents user U1 delivering post P2 once. Positions that appear in the same row have a co-occurrence of delivery of 1. The statistics are totally available, and the delivery co-occurrence times of the position P1 and the { P2, P3, P4, P5} are {2,1,1,3}, respectively.

Second, historical browsing data is collected. In some scenes, the number of resume deliveries received by a position (e.g., a new position) may not be large, at this time, position browsing records of a plurality of users need to be acquired, the times of browsing the position by the users also include, and browsing co-occurrence times of the target position and each other position are respectively calculated according to the position browsing records. For example, the acquired job browsing records of a plurality of users are shown in table 2:

TABLE 2 Job browsing record sheet

	P1	P2	P3	P4	P5
						U1		1	2
U2	2	1	2		4
						U3	1	1	5
U4	6			1	1

Like table 1, U represents a user and P represents a job, then each row in table 2 represents a job browsing record for each user, each column represents a record that each job was browsed, and the numerical value represents the number of browsing, e.g., the number "1" in the third column of the second row represents user U1 browses job P2 once. For example, in row 3 (U2), the co-occurrence value of P1 and P2 is 1, and the co-occurrence value of P1 and P5 is 2, i.e., the smaller value on the corresponding column is taken as the browsing co-occurrence value. The statistical total table shows that the browsing co-occurrence times of P1 and { P2, P3, P4, P5} are {2,2,1,4}, respectively.

It should be noted that, since obtaining the user behavior log (such as a delivery record, a browsing record, and the like) belongs to a common algorithm in the field, description is not expanded here, and any method for obtaining the user behavior log may be combined with the embodiment of the present solution to obtain the position recommendation method described in the present solution.

And finally, respectively calculating the co-occurrence times of the target position and each of other positions according to the delivery co-occurrence times and the browsing co-occurrence times. According to the embodiment of the invention, when the co-occurrence times are calculated, different weights are respectively given to the delivery co-occurrence times and the browsing co-occurrence times, namely, the delivery co-occurrence times and the browsing co-occurrence times are respectively multiplied by a first coefficient and a second coefficient to obtain the treated delivery co-occurrence times and the browsing co-occurrence times, and then the treated delivery co-occurrence times and the browsing co-occurrence times are added to obtain the co-occurrence times. Alternatively, 1 browsing co-occurrence is calculated as 0.2 delivery co-occurrences, then the first coefficient takes 1 and the second coefficient takes 0.2.

Continuing with tables 1 and 2 as examples, assuming that P1 is the target position, the co-occurrence frequency of P1 and { P2, P3, P4, P5} is {2+0.2 × 2,1+0.2 × 1,3+0.2 × 4} - {2.4,1.4,1.2,3.8 }.

Subsequently, in step S220, a first number of positions are sequentially selected from the highest to the lowest co-occurrence numerical values, and a recommended candidate set of the target positions is generated. Continuing with the above table 1 and table 2 as an example, if the first quantity is 3, the recommended candidate set of the target position P1 is { P5, P2, P3}, in turn.

Subsequently, in step S230, on the basis of the number of co-occurrences, consideration of text similarity is added, and the text similarity between the target position and each position in the recommendation candidate set is calculated. And further filtering the positions in the recommendation candidate set according to the text similarity to obtain a more accurate recommendation result.

Specifically, a feature vector of each position is calculated, wherein the feature vector is used for describing the probability distribution of the position on each topic; and calculating the cosine similarity of the included angle between the feature vector of the target position and the feature vector of each position in the recommended candidate set, and taking the cosine similarity as the text similarity.

According to the embodiment of the invention, a feature vector of each position is generated by adopting a document theme generation model (namely, an LDA (latent dirichletailocation) model, the LDA model introduces the concept of 'theme', each word in the document is considered to belong to a specific theme, one document covers a plurality of themes, and the weight of each theme is different, so that the document can be represented as the probability distribution on the themes, namely, doc ═ t (tw ═₁,tw₂,…tw_K) Wherein, tw_iRepresenting the probability of the document being on topic i,

in the embodiment, each position is represented as a K-dimensional feature vector through the LDA model, and is used to describe the probability distribution of the position on K subjects.

According to one implementation, the method 200 further includes the step of training the generative document topic generation model (not shown). The specific method comprises the following steps:

in the first step, 500 ten thousand position details (i.e., 500 ten thousand documents) are selected from a database of a recruitment website as training data, i.e., training samples. Meanwhile, the theme set of the training sample is determined according to the theme (such as "back-end development", "database maintenance", "administration", and the like) covered by each position in the training sample, and generally, the number of the "theme" should be a little more than that of the real theme, so as to improve the accuracy of the algorithm. Optionally, the total number of topics K is taken as 150.

And secondly, randomly assigning a theme to each word in each position in the training sample as an initial theme.

And thirdly, scanning a training sample, and Sampling by using a Gibbs Sampling method (Gibbs Sampling) to obtain the probability that each word belongs to each topic.

Fourth, the sampling step (i.e., step three) is repeated for each roundIteratively recalculating the probability p (z) that each word w belongs to each topic z_k|w_i) Until the probability tends to be stable, the document theme generation model is obtained.

Through training, the probability that each word belongs to each theme is obtained, and when a position comes, the probability distribution of the position on each theme, namely the feature vector of the position, can be obtained through simple summation and summary according to the theme to which each word in the position belongs.

It should be noted that the LDA model is a relatively common document theme generation model, and further details about the LDA model are not described herein. Any processing of the LDA model may be combined with embodiments of the method to arrive at the job recommendations of the present invention.

Then, in step S240, the recommended positions whose calculated text similarity values are not greater than the threshold value are filtered, and only the positions whose text similarity values are greater than the threshold value are selected for recommendation. Thus, when the user browses the page of the target position, the position recommendation result is presented on the meeting page (e.g., below the page).

According to the position recommendation scheme, the relevance between the target position and other candidate positions is described by using the co-occurrence times, specifically, the problem that the coverage rate is low due to too little delivery data is considered, browsing data is added to expand the data volume, the co-occurrence times are calculated together according to the delivery co-occurrence times and the browsing co-occurrence times among the positions, and the problem of data sparseness is solved to a great extent. In addition, text similarity between positions is calculated by using an LDA model, and a recommended candidate set obtained by co-occurrence times is filtered once in the aspect of text similarity, so that the recommendation accuracy is further improved.

By the job recommendation scheme, the job can be more accurately recommended to the user, so that the speed of finding a proper job by a job seeker is increased, each job can be reasonably displayed and recommended, and the satisfaction of both the employing and recruiting parties is improved.

In the description provided herein, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be interpreted as reflecting an intention that: that the invention as claimed requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.

Those skilled in the art will appreciate that the modules or units or components of the devices in the examples disclosed herein may be arranged in a device as described in this embodiment or alternatively may be located in one or more devices different from the devices in this example. The modules in the foregoing examples may be combined into one module or may be further divided into multiple sub-modules.

Those skilled in the art will appreciate that the modules in the device in an embodiment may be adaptively changed and disposed in one or more devices different from the embodiment. The modules or units or components of the embodiments may be combined into one module or unit or component, and furthermore they may be divided into a plurality of sub-modules or sub-units or sub-components. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or elements of any method or apparatus so disclosed, may be combined in any combination, except combinations where at least some of such features and/or processes or elements are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.

Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features included in other embodiments, rather than other features, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments may be used in any combination.

The various techniques described herein may be implemented in connection with hardware or software or, alternatively, with a combination of both. Thus, the methods and apparatus of the present invention, or certain aspects or portions thereof, may take the form of program code (i.e., instructions) embodied in tangible media, such as floppy diskettes, CD-ROMs, hard drives, or any other machine-readable storage medium, wherein, when the program is loaded into and executed by a machine, such as a computer, the machine becomes an apparatus for practicing the invention.

In the case of program code execution on programmable computers, the mobile terminal generally includes a processor, a storage medium readable by the processor (including volatile and non-volatile memory and/or storage elements), at least one input device, and at least one output device. Wherein the memory is configured to store program code; the processor is configured to execute the position recommendation method of the present invention according to instructions in the program code stored in the memory.

Furthermore, some of the described embodiments are described herein as a method or combination of method elements that can be performed by a processor of a computer system or by other means of performing the described functions. A processor having the necessary instructions for carrying out the method or method elements thus forms a means for carrying out the method or method elements. Further, the elements of the apparatus embodiments described herein are examples of the following apparatus: the apparatus is used to implement the functions performed by the elements for the purpose of carrying out the invention.

As used herein, unless otherwise specified the use of the ordinal adjectives "first", "second", "third", etc., to describe a common object, merely indicate that different instances of like objects are being referred to, and are not intended to imply that the objects so described must be in a given sequence, either temporally, spatially, in ranking, or in any other manner.

While the invention has been described with respect to a limited number of embodiments, those skilled in the art, having benefit of this description, will appreciate that other embodiments can be devised which do not depart from the scope of the invention as described herein. Furthermore, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes, and may not have been selected to delineate or circumscribe the inventive subject matter. Accordingly, many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the appended claims. The present invention has been disclosed in an illustrative rather than a restrictive sense, and the scope of the present invention is defined by the appended claims.

Claims

1. A method of job recommendation for recommending a job associated with a target job, the method being performed in a computing device, the method comprising the steps of:

calculating the co-occurrence times of the target position and other positions according to the behavior logs of the users;

sequentially selecting a first number of positions according to the sequence of the co-occurrence numerical values from high to low to generate a recommendation candidate set of target positions;

respectively calculating the text similarity of the target position and each position in the recommendation candidate set; and

selecting positions with text similarity values larger than a threshold value for recommendation;

wherein the step of calculating the text similarity of the target position and each position in the recommendation candidate set respectively comprises:

calculating a feature vector of each position, wherein the feature vector is used for describing the probability distribution of the position on each topic, and the method comprises the following steps: generating a feature vector of each position by using a document theme generating model, and training to generate a document theme generating model;

the training of the generated document theme generation model comprises the following steps:

selecting a plurality of position details from a database of a recruitment website as training samples, and randomly assigning a theme to each word in each position in the training samples as an initial theme;

scanning a training sample, and sampling by using a Gibbs sampling method to obtain the probability that each word belongs to each theme; and

repeating the sampling step until the probability tends to be stable to obtain a document theme generation model;

calculating the cosine similarity of an included angle between the feature vector of the target position and the feature vector of each position in the recommended candidate set as the text similarity;

the step of calculating the co-occurrence times of the target position and other positions according to the behavior logs of the plurality of users comprises the following steps:

acquiring post delivery records of a plurality of users, and respectively calculating the delivery co-occurrence times of a target post and each of other posts according to the post delivery records;

acquiring the position browsing records of a plurality of users, and respectively calculating the browsing co-occurrence times of the target position and each of other positions according to the position browsing records; and respectively calculating the co-occurrence times of the target position and each other position according to the delivery co-occurrence times and the browsing co-occurrence times.

2. The method of claim 1, wherein the step of calculating the number of co-occurrences of the target position and each of the other positions according to the number of delivery co-occurrences and the number of browsing co-occurrences comprises:

multiplying the delivery co-occurrence times by a first coefficient to obtain the processed delivery co-occurrence times, and multiplying the browsing co-occurrence times by a second coefficient to obtain the processed browsing co-occurrence times;

and adding the processed delivering co-occurrence times and the processed browsing co-occurrence times to obtain the co-occurrence times.

3. The method of claim 1, wherein the step of training the document topic generation model further comprises:

and determining a theme set of the training sample according to the theme covered by each position in the training sample.

4. The method of claim 2, wherein the first coefficient is 1 and the second coefficient is 0.2.

5. A computing device, comprising:

one or more processors;

a memory; and

one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs comprising instructions for performing any of the methods of claims 1-4.

6. A computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform any of the methods of claims 1-4.