JP2013008197A - Commodity recommendation processing method, device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To recommend a commodity which corresponds to present proficiency of a user and by which improvement of the proficiency in future is expected.SOLUTION: In a teaching material recommendation processing server SV, expert log data transmitted from a terminal which is used by an expert is received to be accumulated in an expert log database 131, data showing the proficiency of the user at the present point from a terminal of a non-expert user to be stored in a user proficiency database 132. Then, recommendation degrees of all teaching materials to the non-expert user are calculated by a teaching material recommendation processing part 123 based on the accumulated expert log data and the present proficiency of the non-expert user, it is considered that as a teaching material of which the proficiency of the expert when the teaching material is purchased is close to the present proficiency of the user, and the expert can actually improves the proficiency by purchasing the teaching material is the teaching material deserving recommendation to the user more, and recommends the teaching material.

Description

  The present invention relates to a product recommendation processing method, device, and program for recommending a product according to the skill level to a user.

  In educational institutions such as universities and various training institutions, various measures for improving the level of learners or trainees are being studied. As one example, collaborative filtering is used, for example, the degree of similarity in the number of browsing times for the learning material content between learner A and other learners B and C is obtained, and the number of times the common learning material content is browsed based on the result. A technique has been proposed in which learner B or C is discovered as a learner having a similar learning pattern, and teaching material effective for overcoming weakness is taught from learner B or C to learner A (for example, non-patent literature) 1 or 2).

David Goldberg, David Nichols, Brian M. Oki, Douglas Terry, “Using collaborative filtering to weave an information tapestry”, Communications of the ACM-Special issue on information filtering, 35 (12), pp.61-70, 1992. Yuma Tsutsume, Yuji Wada, Kenichi Toi, “Implementation of Recommendation Technology in e-Learning-Extraction of Teaching Materials Contributing to Overcoming Disabilities”, Information Science and Technology Forum (FIT2009), pp. 615-618, 2009.

  However, there is a possibility that a group of people who have purchased or browsed similar learning materials may include people who have different learning levels (skill levels) from those who recommend the learning materials. For this reason, in the method described in Non-Patent Document 1 or 2, that is, the method of finding learners with similar browsing materials by collaborative filtering and recommending the materials used by the learners, how many recommended materials are It is unclear whether it is for people with proficiency levels, and it is not always possible to recommend an appropriate level of teaching material. Some teaching materials set the target skill level. However, it is unclear whether the teaching material is effective in improving the level.

  The present invention has been made paying attention to the above circumstances, and its object is to recommend a product recommendation processing method that can recommend a product that corresponds to the current skill level of the user and can be expected to improve in the future. It is to provide an apparatus and a program.

  In order to achieve the above object, one aspect of the present invention is that information representing purchase or use history of a product by the skilled worker from the terminal used by the skilled worker, and the skilled worker at the time of purchasing or using the product. Expert log information including at least information indicating the skill level of the expert is received and stored in the storage medium. When information representing the current skill level of the unskilled person is received from the terminal used by the unskilled person, the skilled person log information stored in the storage medium and the received unskilled person For each product included in the expert log information based on the information indicating the current skill level of the user, the skill level of the expert and the current skill level of the non-skilled person at the time of purchase or use of the product And the change in the skill level of the expert after the purchase or use of the product is calculated. Then, based on the calculation result of the similarity and the calculation result of the skill level change, a value that increases as the similarity degree is high and the skill level change is large is used as a recommendation degree of the product to the non-experts. The product recommended for the unskilled person is selected based on the recommendation degree calculated for each product.

により計算し、
上記熟練者の熟練度の変化を計算する過程では、商品Ｂを購入又は使用後の熟練者ｉの熟練度の変化を

により計算し、
上記商品の推薦度を計算する過程では、Ｅ_B を商品Ｂを購入又は使用した熟練者の集合、ｎ_B を商品Ｂを購入又は使用した熟練者の人数、ｗを各要素に付与される重みとしたとき、商品Ｂの推薦度rec_B を

により計算する。 Specifically, in the process of calculating the degree of similarity, proficiency number when the m _B skill has purchased a product _B, l _i [m B] a m _B th skill of skill i, l _{When u} is the skill level of the unskilled person, the similarity sim _{i_B} between the skill level of the skill person i when the product B is purchased and the current skill level of the unskilled person is

Calculated by
In the process of calculating the change in the skill level of the expert, the change in the skill level of the expert i after purchasing or using the product B is calculated.

Calculated by
In the process of calculating the recommendation degree of the product, the weight applied set skill who purchase or use a product B to E _B, those skilled in number purchased or used a product B a n _B, the w to each element And the recommendation level rec _B of product B

Calculate according to

  Therefore, among the products purchased by the skilled person in the past, the skill level of the skilled person at the time of purchase is similar to the current skill level of the unskilled person and the improvement rate of the skill level of the skilled person after the purchase is high. Is recommended as a product suitable for the unskilled person. For this reason, it is possible for the unskilled user to receive a recommendation of a product having an appropriate difficulty level corresponding to the current skill level and having a high possibility of future growth like the skilled person.

により計算するようにしてもよい。
このようにすると、非熟練者に対する商品の推薦度を、熟練者間平均の重み付き和だけでなく、熟練者間の和の重み付き和及び熟練者間の積の形も考慮して計算することが可能となる。 In addition, in the process of calculating the recommendation level of the product, the recommendation level rec _B of the product _B for the unskilled person is considered in consideration of the weighted sum of the sum between experts and the form of the product between the experts.

You may make it calculate by.
In this way, the degree of product recommendation for non-experts is calculated in consideration of not only the weighted sum of averages among experts, but also the weighted sum of sums between experts and the form of products between experts. It becomes possible.

One aspect of the present invention is also characterized by comprising the following various aspects.
In the first aspect, in the process of calculating the recommendation degree of the product, when the expert log information includes information indicating the usage time of the product after purchase or use by the expert, the similarity degree Based on the calculation result of the above, the calculation result of the change in the skill level, and the use time of the product after purchase or use, the similarity is high and the change in the skill level is large. A value that increases as the usage time of the product increases is calculated as the degree of recommendation of the product for the unskilled person.
In this way, "the skill level of the skilled person when purchasing the product is close to the current skill level of the unskilled person", "the expert actually improved the skill level by purchasing the product" In addition to the two conditions, it is possible to calculate the recommendation level of the product by further considering the condition that “the product is used by the expert for a long time”, thereby further improving the calculation accuracy of the recommendation level. .

In the second aspect, in the process of calculating the recommendation degree of the product, when the expert log information includes information indicating the evaluation of the product by the expert, the similarity calculation result, Based on the calculation result of the skill level change and the information indicating the evaluation of the product, a value that increases as the similarity degree is high, the skill level change is large, and the product evaluation by the expert is high. The degree of recommendation of the product for the unskilled person is calculated.
In this way, "the skill level of the skilled person when purchasing the product is close to the current skill level of the unskilled person", "the expert actually improved the skill level by purchasing the product" In addition to the two conditions, it is possible to calculate the recommendation level of the product by further considering the condition that “the product is highly evaluated by the expert”, and thereby the calculation accuracy of the recommendation level can be further increased.

In the third aspect, in the process of calculating the recommendation degree of the product, the number of skilled persons who have purchased or used the product is calculated for each product based on the expert log information, and the similarity is calculated. Based on the result, the calculation result of the change in the skill level, and the calculation result of the number of skilled persons who purchased or used the product, the similarity is high and the change in the skill level is large. A value that increases as the number of used skilled workers increases is calculated as the degree of recommendation of the product for the unskilled person.
In this way, "the skill level of the skilled person when purchasing the product is close to the current skill level of the unskilled person", "the expert actually improved the skill level by purchasing the product" In addition to the two conditions, it is possible to calculate the recommendation level of the product by further considering the condition that “the product was purchased or used by many experts”, thereby further improving the calculation accuracy of the recommendation level. be able to.

  That is, according to the present invention, it is possible to provide a product recommendation processing method, apparatus, and program that can recommend a product that corresponds to the current skill level of the user and that can be expected to improve in the future.

1 is a schematic configuration diagram of a system including a product recommendation processing device according to an embodiment of the present invention. The block diagram which shows the function structure of the teaching material recommendation process server which is a goods recommendation processing apparatus which concerns on one Embodiment of this invention. The figure which shows the flow of the information in the system shown in FIG. The flowchart which shows the process sequence and process content of the teaching material recommendation process server shown in FIG. The figure which shows the 1st example of the expert log data provided from the expert. The figure which shows the 2nd example of the expert log data provided from the expert. The figure which shows the 3rd example of the expert log data provided from the expert. The figure which showed an example of the recommendation result of the teaching material by the teaching material recommendation process server shown in FIG. 2 in contrast with the recommendation result by a prior art. The figure which shows the characteristic of the teaching material recommended based on the recommendation degree calculated by this embodiment.

Embodiments according to the present invention will be described below with reference to the drawings.
[One Embodiment]
(Constitution)
FIG. 1 is a schematic configuration diagram of a system for realizing a product recommendation processing method according to an embodiment of the present invention, and provides a service for recommending teaching materials as products.
This system includes a teaching material recommendation processing server SV having a function as a product recommendation processing device, and this teaching material recommendation processing server SV, terminals MS1 to MSn used by experts and terminals UT1 to UTm used by non-skilled users. Can be communicated with each other via the communication network NW. The communication network NW includes, for example, an IP (Internet Protocol) network represented by the Internet and a plurality of access networks for accessing the IP network. As the access network, for example, a LAN (Local Area Network), a wireless LAN, a mobile phone network, a wired phone network, and a CATV (Cable Television) network are used.

  The terminals MS1 to MSn used by skilled users and the terminals UT1 to UTm used by non-expert users are mobile phones, smart phones, tablet terminals, PDAs (Personal Digital Assistants), notebook personal computers, etc. The mobile terminal is provided with a web access function such as a browser and an e-mail transmission / reception function.

  The expert terminals MS1 to MSn have a function of acquiring and transmitting expert log data. This function has an automatic acquisition mode or a manual acquisition mode as a mode for acquiring expert log data. In the automatic acquisition mode, when an expert purchases teaching materials on an online site and uses this teaching material to access an e-learning site and learn, attribute information such as the name of the purchased teaching material is acquired from the online site. Then, information indicating the proficiency level of the expert, the usage time of the teaching material, and the evaluation is acquired from the e-learning site. On the other hand, in the manual acquisition mode, the attribute information of the purchased teaching material input by the expert himself, the skill level of the expert, the usage time and evaluation of the purchased teaching material are taken from the input device.

  Further, the skilled person log data acquisition / transmission function has acquired the attribute information of the purchased teaching material, the proficiency level of the skilled worker, the usage time of the purchased teaching material, and the information acquired in the automatic acquisition mode or the manual acquisition mode. The user log data is transmitted to the teaching material recommendation processing server SV using the browser or e-mail.

The user terminals UT1 to UTm have a function of acquiring and transmitting information representing the user's proficiency level. This function also has an automatic acquisition mode and a manual acquisition mode, similar to the terminals MS1 to MSn of the expert. In the automatic acquisition mode, for example, when a user takes a course using an e-learning site, information indicating his / her proficiency level is acquired from the site. In the manual acquisition mode, information indicating the level of proficiency input by the user himself / herself is acquired from the input device.
The user skill level acquisition / transmission function transmits information indicating the user's proficiency level acquired in the automatic acquisition mode or the manual acquisition mode to the teaching material recommendation processing server SV using the browser or e-mail.

By the way, the teaching material recommendation processing server SV is composed of, for example, a Web server and is configured as follows. FIG. 2 is a block diagram showing the functional configuration.
That is, the teaching material recommendation processing server SV includes a communication interface unit 11, a control unit 12, and a storage unit 13.

  Under the control of the control unit 12, the communication interface unit 11 is adapted to the expert log data transmitted from the terminals MS <b> 1 to MSn used by the skilled person and the terminals UT <b> 1 to UTm used by the unskilled user and the user's proficiency. A process of receiving information indicating the degree via the communication network NW is performed. Moreover, the process which transmits the information showing the recommendation result produced | generated by the control unit 12 to the terminal UT1-UTm of an unskilled person via the communication network NW is performed.

  The storage unit 13 uses a randomly accessible non-volatile memory using, for example, an HDD (Hard Disc Drive) or an SSD (Solid State Drive), and is used as a storage unit necessary for realizing the embodiment of the present invention. A skilled person log database 131 and a user skill level database 132 are provided.

  The expert log database 131 is used for storing expert log data collected from the terminals MS1 to MSn of the expert. The user skill level database 132 is used to store data representing user skill levels collected from the terminals UT1 to UTm used by non-skilled users.

  The control unit 12 includes a central processing unit (CPU) as a core. As control functions necessary for carrying out the embodiment of the present invention, an expert log acquisition control unit 121, a user skill level, The acquisition control unit 122, the teaching material recommendation processing unit 123, and the recommendation information transmission control unit 124 are provided. All of these control functions are realized by causing the CPU to execute an application program stored in a program memory (not shown) in the storage unit 30.

  The expert log acquisition controller 121 receives the expert log data via the communication interface unit 11 every time expert log data is transmitted from the terminals MS1 to MSn used by the expert. Then, the received expert log data is stored in the expert log database 131 in association with the identification information (expert ID) of the expert who is the transmission source.

  The user skill level acquisition control unit 122 receives the user skill level data via the communication interface unit 11 every time data representing the skill level of the user is transmitted from the terminals UT1 to UTm used by the non-expert user. To do. Then, the received user skill level data is stored in the user skill level database 132 in association with identification information (user ID) of the transmission source user.

  When attempting to recommend a learning material to a user, the learning material recommendation processing unit 123 reads data representing the skill level of the user from the user skill level database 132 and logs all skilled workers from the expert log database 131. Read data. Then, based on the read information indicating the skill level of the user and the log data of all the skilled persons, the recommendation level of all teaching materials for the certain user is calculated, and based on the result, the user is notified. The process of selecting the most appropriate teaching material for recommendation is performed.

There are the following three types of recommendation level calculation and selection methods.
1st method "[1] The skill level of the expert when purchasing the teaching material is close to the skill level of the user", "[2] The expert actually improved the skill level by purchasing the teaching material" That is, a teaching material that satisfies the above two conditions is considered to be a teaching material worthy of recommendation to the user.

Second Method In addition to the above conditions [1] and [2], “[3] Highly used time for teaching materials by experts”, “[4] Highly evaluated teaching materials by experts”, “[5] A method in which three conditions of “purchased or used by many experts” are added, and a teaching material satisfying these conditions is considered to be a teaching material worthy of recommendation to the user.

The third method Focus on the skill level when using the teaching material instead of the skill level when purchasing the teaching material, and "[1] 'The skill level of the skilled person when using the teaching material is close to the skill level of the user", "[ 2] 'The expert actually improved the skill level by using the teaching materials'. Furthermore, the teaching materials satisfying the above conditions [3], [4] and [5] A method that considers teaching materials more worthy of recommendation.

  The recommendation information transmission control unit 124 performs processing for transmitting information related to the learning material selected by the learning material recommendation processing unit 123 from the communication interface unit 11 to the terminal of the user to be notified.

(Operation)
Next, the operation of the system configured as described above and the teaching material recommendation processing server SV will be described. FIG. 3 is a diagram showing the information flow of the system.
(1) Expert log data collection process In the standby state, the control unit 12 activates the expert log acquisition control unit 121, and under the control of the expert log acquisition control unit 21, a previously registered expert The skilled person log acquisition request is transmitted from the communication interface unit 11 to the terminals MS1 to MSn to be used at different timings. And the expert log data transmitted in step S1 from each terminal MS1-MSn in response to this request is received via the communication interface unit 11. Then, the received expert log data is stored in the expert log database 131 in association with the ID of the sender expert. The expert log database 131 stores log data for the past approximately one month when it is short and for the past approximately ten years when it is long.

  The expert log data includes, as elements, attribute information indicating the title of the purchased teaching material, the skill level of the expert at the time of purchasing the teaching material, the usage time after purchasing the teaching material, and the evaluation of the teaching material. FIGS. 5A and 5B show expert log data of skilled workers M1 and M2 as an example. In this example, the TOEIC (registered trademark) score is used as the skill level.

(2) Acquisition processing of skill level data of unskilled user It is assumed that a teaching material introduction request is received from the terminal UT1 of the unskilled user U1 who has been registered for use with this system. Then, the control unit 12 first transmits a skill level transmission request to the terminal UT1 of the requesting non-skilled user U1 by the user skill level acquisition control unit 122. When data indicating the skill level of the user U1 at the present time is transmitted from the terminal UT1 in response to this request (step S2), the user skill level acquisition control unit 122 transmits the data indicating the skill level to the communication interface unit 11. Receive via. Then, the received data indicating the skill level is stored in the user skill level database 132 in association with the user ID of the non-skilled user who is the transmission source. As the data representing the skill level, TOEIC scores are used as in the expert log data described above.

(3) Teaching material recommendation processing When the skill level data acquisition processing of the non-skilled user U1 as the request source is completed, the control unit 12 starts up the teaching material recommendation processing portion 123, and thereafter, under the control of the teaching material recommendation processing portion 123. The teaching material recommendation process is executed as follows (step S3 in FIG. 3). FIG. 4 is a flowchart showing the processing procedure and processing contents.

  First, in step S31, the learning material recommendation processing unit 123 reads data representing the skill level of the introduction requesting user from the user skill level database 132. Subsequently, in step S32, expert log data including the same type of skill level data as the user skill level data is read from the expert log database 131. For example, if the user's skill level is a TOEIC score, expert log data including the TOEIC score is read. Then, in step S33, it is determined whether or not reading of all corresponding expert log data has been completed. If not, the process returns to step S32 to continue reading expert log data.

  When the reading of all the expert log data is completed, the teaching material recommendation processing unit 123 next, in step S34, based on the read user skill level data and expert log data, all the teaching material for the requesting user. The degree of recommendation is calculated by selectively using the following three methods.

(3-1) First Method As described above, the first method is “[1] The skill level of the skilled person when purchasing the teaching material is close to the skill level of the user”, “[2] Teaching material The teaching material that satisfies the two conditions that “the expert has actually improved the skill level by purchasing“ is considered to be a teaching material that deserves more recommendation to the user.

ただし、ｍ_B は熟練者が教材Ｂを購入したときの熟練度番号を示し、ｌ_i [ｍ_B] は熟練者ｉのｍ_B 番目の熟練度、ｌ_u はユーザの熟練度をそれぞれ示す。 In this first method, the elements of the expert log data used for the calculation are only attribute information and skill level of the purchased teaching material as shown in FIG.
Now, focusing on the learning material B, a case where the recommendation level of the learning material B for the user is calculated will be described as an example.
The condition shown in [1], that is, the similarity sim _{i_B} between the skill level of the expert i at the time of purchasing the teaching material B and the skill level of the user is calculated by the following equation.

Here, m _B indicates the skill level number when the expert purchases the teaching material B, l _i [m _B ] indicates the m _B- th skill level of the expert i, and l _u indicates the skill level of the user.

ただし、低いほど良い熟練度（例えばゴルフスコア）の場合は、マイナスをかける。 The condition shown in [2] is calculated by the following equation as the recommendation degree rec _{i_B} of the teaching material B by the expert i.

However, if the skill level is better as it is lower (for example, golf score), a minus is applied.

ただし、Ｅ_B は教材Ｂを購入した熟練者の集合を、ｎ_B は教材Ｂを購入した熟練者の人数をそれぞれ示す。また、ｗは各要素に付与される重みである。 Recommendation level when combining the conditions of [1] and [2], that is, recommendation of teaching material B to the user when combining the calculation formulas of [1] and [2] and considering log data of multiple experts The degree rec _B is expressed as follows.

Here, E _B represents a set of skilled persons who purchased the teaching material B, and n _B represents the number of skilled persons who purchased the teaching material B. W is a weight given to each element.

のように表される。
上記式(3) において、類似度sim_{i_B} 及び推薦度rec_B が０以下の値になる場合は、代わりに０．１や０．０１などの低い値を割り当てる。 Furthermore, regarding the above formula (1), considering not only the average weighted sum between experts, but also the weighted sum of sums between experts and the form of the product between experts, the recommendation of teaching material B to the user Degree rec _B

It is expressed as
In the above formula (3), when the similarity sim _{i_B} and the recommendation degree rec _B are values of 0 or less, low values such as 0.1 and 0.01 are assigned instead.

(3-2) Second method As described above, in the second method, in addition to the above conditions [1] and [2], “[3] More time spent on teaching materials by experts”, “ [4] Teaching materials are highly evaluated by experts, and “[5] Purchased or used by many skilled people”, and materials that satisfy these conditions are more recommended to users. It is a method that is considered to be teaching material.

In this second method, the elements of the expert log data used for the calculation are attribute information, skill level, usage time, evaluation, and the number of experts who have purchased the teaching material shown in FIG.
Now, as in the case of the first method, a case where attention is paid to the learning material B and the recommendation level of the learning material B for the user is calculated will be described as an example.

ただし、ｔ_{i_j} は熟練者ｉによる教材ｊの使用時間を、ｒ_{i_j} は熟練者ｉによる教材ｊの評価をそれぞれ示している。またＩ_i [ｍ_B] は、熟練者の熟練度がｌ_i [ｍ_B] のときに熟練者が使用した教材の集合を示している。
また上記式において、

は、Ｉ_i [ｍ_B] の中で、教材Ｂが次の熟練度番号への熟練度の向上にどれくらい貢献したかを表している。またｎ_B は、[5] の条件で述べた、教材Ｂを購入した熟練者の人数を示している。 First, the recommendation degree rec2 _{i_B} of the teaching material B by the expert i considering the conditions [3] and [4] is expressed by the following equation.

Here, t _{i_j} indicates the usage time of the teaching material j by the expert i, and r _{i_j} indicates the evaluation of the teaching material j by the expert i. In addition, I _i [m _B ] indicates a set of teaching materials used by the expert when the skill level of the expert is l _i [m _B ].
In the above formula,

Represents how much the teaching material B contributed to the improvement of the skill level to the next skill level number in I _i [m _B ]. N _B indicates the number of skilled persons who have purchased the teaching material B described in the condition of [5].

また、上記式(4) についても、先に述べた式(1) と同様に、熟練者間平均の重み付き和だけではなく、熟練者間の和の重み付き和や、熟練者間の積の形も考慮する。 Then, when the calculation formulas representing the above conditions [1] to [5] are summarized and log data of a plurality of experts are taken into consideration, the recommendation level rec2 _B of the teaching material B by the expert i is calculated by the following formula: .

Also, in the above equation (4), not only the average weighted sum between experts, but also the weighted sum of sums between experts, and the product between experts, as in equation (1) described above. Also consider the form.

ただし、ｐは順番による貢献度を上げる割合を示し、その範囲は例えば０．０１＜ｐ＜１とする。Ｏ_B はＩ_i [ｍ_B] の中で教材Ｂが購入された順番を示している。 Further, regarding the recommendation level rec2 _{i_B} of the learning material B, the following form for improving the recommendation level of the learning material purchased immediately before improving the skill level is considered.

However, p shows the ratio which raises the contribution by order, and the range is set to 0.01 <p <1, for example. O _B indicates the order in which they are purchased materials B in I _i [m _B].

On the other hand, considering the case where the recommendation level is improved as the usage time of the learning material is reduced, the recommendation level I _i [m _B ] of the learning material B is expressed as follows.

(3-3) Third Method As described above, the third method focuses on the proficiency level when using teaching materials instead of the proficiency level when purchasing teaching materials. Satisfies the skill level of the expert of the user ”and“ [2] ′ The skill level was actually improved by using the teaching materials ”and the above [3] In other words, the teaching materials satisfying the conditions of [4] and [5] are regarded as the teaching materials worthy of recommendation to the user.

In the third method, the elements of the expert log data used for the calculation are attribute information of the purchased teaching material, the skill level at the time of use, the usage time, the evaluation, and the number of skilled workers who purchased the teaching material as shown in FIG. Become.
Now, as in the case of the first method, a case where attention is paid to the learning material B and the recommendation level of the learning material B for the user is calculated will be described as an example.

ただし、ｌ_i [ｍ_B][q] は熟練者ｉが教材Ｂをｑ回目に使用したときの熟練度、Ｉ_i [ｍ_B][q]はｌ_i [ｍ_B][q]のときに使用した教材の集合をそれぞれ表している。 The degree of recommendation rec3 _{i_B} of the teaching material B by the expert i when paying attention to the skill level when using the teaching material is calculated by the following equation.

However, l _i [m _B ] [q] is the skill level when expert i uses teaching material B q times, and I _i [m _B ] [q] is l _i [m _B ] [q] Each represents a set of teaching materials used.

そして、教材使用時の熟練度に着目した、ユーザに対する教材Ｂの推薦度rec3_B は、次式により計算される。

なお、以上の説明では、何れの方法も教材Ｂの推薦度を算出する場合を例にとって説明したが、同じ計算を熟練者ログデータに含まれるその他の各教材について繰り返し実行し、最終的に熟練者ログデータに含まれる全ての教材について推薦度を算出する。その算出結果は制御ユニット５内の作業用メモリに保存される。 Further, sim3 _{i_B} is a similarity between the skill level of the expert i and the user when using the teaching material B, paying attention to the skill level when using the teaching material, and is calculated by the following equation.

The recommendation level rec3 _B of the learning material B to the user, focusing on the skill level when using the learning material, is calculated by the following equation.

In the above description, each method has been described with respect to the case where the recommendation level of the teaching material B is calculated as an example. However, the same calculation is repeatedly performed for each of the other teaching materials included in the expert log data, and finally the skill level is determined. The recommendation level is calculated for all the teaching materials included in the person log data. The calculation result is stored in a working memory in the control unit 5.

(4) Selection and transmission of recommended learning materials When the recommendation levels for all the learning materials are calculated in step S34, the control unit 12 reads the calculated recommendation levels for all the learning materials from the work memory and Compare. As a result of the comparison, the learning material (book, video, audio teaching material) having the highest recommendation level is selected, the attribute information of the selected learning material is read from the expert log database 131, and the communication interface unit 11 introduces the introduction request source. To the user terminal UT1.

  In selecting a learning material, a preset number of learning materials may be selected from those having a high recommendation level, and the attribute information may be transmitted to the user terminal with information indicating the recommendation order. In this case, teaching materials whose recommendation level does not satisfy a preset threshold value may be excluded.

(Effect of embodiment)
As described in detail above, in this embodiment, the teaching material recommendation processing server SV receives expert log data transmitted from the terminals MS1 to MSn used by the expert and stores it in the expert log database 131. Data representing the current skill level of the user is received from the expert user terminals UT <b> 1 to UTm and stored in the user skill level database 132. Then, based on the accumulated log data of the skilled person and the current skill level of the unskilled user, the teaching material recommendation processing unit 123 calculates the degree of recommendation of all the teaching materials for the non-skilled user. Based on the calculation result, the teaching material to be recommended is selected and the introduction information is transmitted to the user terminal.
At that time, the recommendation level calculation method is as follows: “[1] The skill level of the expert when purchasing the teaching material is close to the current skill level of the user”, “[2] The teaching material that satisfies the two conditions of “actually improved skill level” is used as a teaching material that deserves more recommendation to the user.

  Therefore, among the teaching materials previously purchased by the expert, the skill level of the expert at the time of purchase is similar to the current skill level of the non-skilled user, and the improvement rate of the skill level of the expert after the purchase is high. The learning material is recommended as the learning material suitable for the unskilled user. For this reason, it is possible for a non-skilled user to receive a recommendation of a teaching material that has an appropriate difficulty level corresponding to the current skill level and has a high possibility of future growth like a skilled person.

  FIG. 9 shows the characteristics of the teaching material recommended based on the recommendation degree calculated by the present embodiment when recommending the teaching material to a non-expert user with a current TOEIC score of 450 points. As is clear from this example, teaching materials that are expected to grow in the future can be recommended to unskilled users. Incidentally, FIG. 8 shows the characteristics of a learning material recommended by a conventional method, that is, a conventional method of finding a learner whose browsing material is similar by collaborative filtering and recommending the learning material used by this learner. Since only the teaching materials used by learners with the same skill level are recommended, the growth of unskilled users is often not expected.

  Further, in this embodiment, the conditions [1] and [2] described above are “[3] Time spent on learning materials by experts”, “[4] High evaluation of teaching materials by experts”, “[5 ] When the recommendation level of a teaching material is calculated with the addition of the three conditions “purchased or used by many experts”, the teaching material satisfying these conditions is considered to be a teaching material worthy of recommendation to the user. I am doing so. Therefore, in addition to the above conditions [1] and [2], the degree of recommendation of the teaching material is calculated in consideration of the above conditions [3], [4] and [5]. Can be increased.

  Furthermore, this embodiment also proposes a method for calculating a recommendation level based on the skill level of an expert when using a teaching material and the skill level actually improved by using the teaching material. If the calculation method based on this proposal is used, the degree of recommendation is calculated based on the usage of the learning material after purchase, so it is possible to perform recommendation processing that reflects the actual effect after the use of the learning material starts. Become.

[Other Embodiments]
The present invention is not limited to the above embodiment. For example, in the embodiment, as a method for calculating the recommendation level of the teaching material, the teaching materials for the unskilled user are added by adding all the conditions [3], [4], and [5] to the conditions [1] and [2]. The recommendation level is calculated, but one or two of [3], [4], and [5] are added to the conditions of [1] and [2], and The recommendation level may be calculated.

  Moreover, although the case where collection of expert log data from each expert terminal MS1 to MSn by the teaching material recommendation processing server SV is performed using a polling method has been described as an example, each expert terminal MS1 to MSn uses a push method. The expert log data may be transmitted, and the teaching material recommendation processing server SV may receive it each time and store it in the expert log database 131.

Furthermore, in the above embodiment, the case where the TOEIC score is applied as the skill level and the teaching material is recommended has been described as an example. However, the teaching material is recommended by applying the scores of various mock tests, golf scores, and the like as the skill level. Even in this case, the present invention can be applied.
In addition, the configuration of the teaching material recommendation processing server and the processing procedure and processing contents thereof, the types and functions of terminals used by experts and non-expert users, the types of products, and the like deviate from the gist of the present invention. Various modifications can be made without departing from the scope.

  In short, the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. Further, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, you may combine suitably the component covering different embodiment.

  SV ... teaching material recommendation processing server, MS1 to MSn ... terminals used by skilled users, UT1 to UTm ... terminals used by unskilled users, NW ... communication network, 11 ... communication interface unit, 12 ... control unit, 13 ... memory Units 131, 131 ... Expert log database 131, 132 ... User skill level database, 121 ... Expert log acquisition control unit, 122 ... User skill level acquisition control unit, 123 ... Teaching material recommendation processing unit, 124 ... Recommendation information transmission control unit .

Claims (8)

Received from the terminal used by the expert is expert log information including at least information indicating the purchase or use history of the product by the expert and information indicating the skill level of the expert at the time of purchase or use of the product. Storing in a storage medium;
A process of receiving information representing the current skill level of the unskilled person from the terminal used by the unskilled person,
Based on the expert log information of the skilled person stored in the storage medium and the information indicating the current skill level of the received non-expert person, for each commodity included in the expert log information, the commodity Calculating a similarity between the skill level of the expert at the time of purchase or use and the current skill level of the non-skilled person;
A process of calculating, for each product included in the expert log information, a change in the skill level of the expert after the purchase or use of the product based on the expert log information of the expert stored in the storage medium When,
Based on the calculation result of the similarity and the change in the skill level, a value that increases as the similarity is high and the change in the skill level is large is calculated as a recommendation degree of the product for the unskilled person The process of
A product recommendation processing method comprising: selecting a product recommended for the unskilled person based on a recommendation degree calculated for each product. In the process of calculating the similarity, m _B is a skill level number when an expert purchases a product B, l _i [m _B ] is the m _B- th skill level of expert i, and l _u is an unskilled person. The degree of similarity sim _{i_B} between the skill level of the expert i at the time of purchasing the product B and the current skill level of the non-skilled person when the product B is purchased

Calculated by
The process of calculating the change in the skill level of the expert is the change in the skill level of the expert i after purchasing or using the product B.

Calculated by
Process of calculating the recommendation degree of the product, the weight applied set skill who purchase or use a product B to E _B, those skilled in number purchased or used a product B a n _B, the w to each element And the recommendation level rec _B of product B

The product recommendation processing method according to claim 1, wherein the product recommendation processing method is calculated by: In the process of calculating the recommendation level of the product, the recommendation level rec _B of the product _B for the unskilled person is considered in consideration of the weighted sum of the sum between experts and the form of the product between the experts.

The product recommendation processing method according to claim 2, further comprising the step of calculating according to claim 1.   The step of calculating the recommendation level of the product includes the calculation result of the similarity when the expert log information includes information indicating the usage time of the product after purchase or use by the expert, Based on the calculation result of the change in skill level and the usage time of the product after purchase or use, the similarity is high, the change in skill level is large, and the use time of the product by the expert is longer 2. The product recommendation processing method according to claim 1, wherein a value to be increased is calculated as a recommendation level of the product for the unskilled person.   The process of calculating the recommendation level of the product includes the calculation result of the similarity and the calculation of the change in the skill level when the expert log information includes information indicating the evaluation of the product by the expert. Based on the result and information representing the evaluation of the product, a value that increases as the similarity is high and the change in the skill level is large, and the product evaluation by the expert is high, The product recommendation processing method according to claim 1, wherein the product recommendation processing is calculated as a recommendation level of the product. The process of calculating the recommendation level of the product is as follows:
Based on the expert log information, a process of calculating the number of experts who have purchased or used the product for each product;
Based on the calculation result of the similarity, the calculation result of the change in the skill level, and the calculation result of the number of skilled persons who have purchased or used the product, the similarity is high and the change in the skill level is large. 2. The product recommendation processing method according to claim 1, further comprising a step of calculating, as the recommendation level of the product for the non-experts, a value that increases as the number of skilled or purchased experts increases. . Received from the terminal used by the expert is expert log information including at least information indicating the purchase or use history of the product by the expert and information indicating the skill level of the expert at the time of purchase or use of the product. Means for storing in a storage medium;
Means for receiving information representing the current skill level of the unskilled person from a terminal used by the unskilled person;
Based on the expert log information of the skilled person stored in the storage medium and the information indicating the current skill level of the received non-expert person, for each commodity included in the expert log information, the commodity Means for calculating the similarity between the skill level of the skilled person at the time of purchase or use and the current skill level of the non-skilled person;
Based on the expert log information of the expert stored in the storage medium, for each product included in the expert log information, means for calculating a change in the skill level of the expert after the purchase or use of the product When,
Based on the calculation result of the similarity and the change in the skill level, a value that increases as the similarity is high and the change in the skill level is large is calculated as a recommendation degree of the product for the unskilled person Means to
A product recommendation processing apparatus comprising: means for selecting a product to be recommended to the unskilled person based on a recommendation degree calculated for each product.   The program which makes the computer with which the said product recommendation processing apparatus performs the process which each means with which the product recommendation processing apparatus of Claim 7 comprises is performed.

Images