KR101723770B1 - Method and system for recommending problems based on player matching technique - Google Patents

Abstract

Provided are a question recommendation method based on a player matching technique and a system thereof. The question recommendation method based on player matching technique according to an embodiment of the present invention can comprise the steps of: checking the level of an inputted player; checking the level of questions based on whether or not the questions were solved; determining a winning probability for the questions of the player according to the level of the player and the level of the questions; listing the questions in order of the winning probability close to 50%; and providing the designated number of questions among the listed questions to the player.

Description

METHOD AND SYSTEM FOR RECOMMENDING PROBLEMS BASED ON PLAYER MATCHING TECHNIQUE

The present invention relates to problem recommendations, and more particularly, to problem recommendation systems based on player matching techniques.

In the computer programming domain, there are many conventions that measure programmers' programming skills. These competitions are supported by prominent IT companies, many programmers dream of winning these competitions, and IT companies offer not only prize money but also job opportunities.

For this reason, there are a number of Web services that provide programmers with exercises to prepare for the competition. However, these problem provision services do not efficiently provide adequate problems according to the ability of programmers due to too many conferences and problems and participants.

If a Web service has a system that recommends programmers the right level of difficulty, programmers will be able to improve their capabilities more effectively.

Patent Document 10-1089163 discloses a content database in which a plurality of contents and difficulty information of the contents are stored; An achievement calculator for calculating an achievement score of a content stored in the content database from a learner; A similarity calculation unit for calculating a similarity between the content learned by the learner and the content stored in the content database; And a degree of difficulty of the content stored in the content database, an achievement degree of the learner measured by the achievement degree calculation unit, and the degree of similarity information calculated by the degree of similarity calculation unit, And a content recommending unit that selects a succeeding learning content to be learned by the learner among the content stored in the content database so that the learning content is selected and the degree of difficulty and the degree of similarity of the posterior learning content become lower as the achievement degree is lower Discloses that the learner determines the trailing learning contents based on the score of the contents solved and the difficulty of the contents by providing the contents providing apparatus.

However, the problems presented to the player are determined by the degree of difficulty already determined, the player only determines the player's rating according to the score of the problem solved, and providing the post-learning determined to be similar to the determined grade, There is a limitation in providing a problem of degree of difficulty suitable for the skill of the player.

KR 10-2011-0134314 (Published patent)

As described above, there is a need for a problem recommendation method and a system for providing a problem of difficulty suitable for a player's actual skill in the preparation of a competition for a player.

According to various embodiments of the present invention, it is possible to provide a method and system for recommending a problem appropriate to a player's rating by comparing the rating of each of the problems with the rating of the player.

According to various embodiments of the present invention, there is provided a method of comparing a rating of a player with a rating of a player based on a rating of each rating of a problem, Can be provided.

According to an embodiment of the present invention, a problem recommendation method based on a player matching technique includes: checking a rating of an input player; Confirming the rating of the problems based on whether or not to solve the problems; Determining a winning probability for the player's problems according to the rating of the player and the rating of the problems; Enumerating the problems according to an order in which the victory probability is close to a specified value; And providing the player with a designated number of the listed problems.

According to various embodiments, the rating of the player is determined based on information of the player input, and when the rating of the player is not specified, at least one of the problems corresponding to two or more difficulties is provided Can be determined based on the test problem.

According to various embodiments, the grade of the problems may include at least one of the level of the problems, the level of the competition including the problems.

According to various embodiments, a problem recommendation method based on a player matching technique comprises: determining a probability that the problems are at a level suitable for the player based on the rating of the problems and the rating of the player; And providing the player with a designated number in the order of higher probability of the appropriate level.

According to various embodiments, whether or not to overcome the problems in the database is determined by determining that the player has won when the player inputs a correct answer to the question, and if the player has input a wrong answer to the question, You can decide to be defeated.

 According to various embodiments, determining the victory probability for the player's problems based on the rating of the player and the rating of the problems may include comparing the rating of the player and the rating of the problems with at least one player matching technique And the like.

According to various embodiments, a problem recommendation method based on a player matching technique comprises: updating the win / loss of the player against the provided problem to the database; And adjusting the rating of the player based on the win / loss of the player.

According to one embodiment of the present invention, a problem recommendation system based on a player matching technique comprises: a memory including a database of problems; A player input unit for receiving information from the player; A recommendation problem providing unit for providing problems to the player; And a processor coupled to the memory, the player input unit, and the recommendation problem providing unit, wherein the processor confirms a rating of a player input through the player input unit, Determining a winning probability for the player's problems based on the rating of the player and the rating of the problems and comparing the winning probabilities to the problems listed And process the designated number of the listed problems to provide the player with the recommended problem providing unit.

According to various embodiments, the processor may determine the player's rating based on information of the player input, and if the player's rating is not specified, at least one of the problems corresponding to two or more difficulties The player can determine the rating of the player based on the test questions provided.

According to various embodiments, the processor may include at least one of a level of the problems, a level of competition including the problems, in the class of the problems.

According to various embodiments, the processor is further configured to determine a probability that the problems are at a level suitable for the player based on the rating of the problems and the rating of the player, To be provided to the player.

According to various embodiments, the processor may determine that the player has won when the player enters a correct answer to the question, and may determine that the player is defeated if the player has entered an incorrect answer.

According to various embodiments, the processor may apply at least a portion of the rating of the player and a class of the problems to at least one player matching technique to determine a win probability for the player's problems.

According to various embodiments, the processor may process the player to win or lose against the provided problem to the database, and adjust the player's rating based on win or loss of the player.

According to various embodiments of the present invention, the player's skill can be improved by comparing the rating of each of the problems with the rating of the player and recommending a problem that is appropriate to the rating of the player.

According to various embodiments of the present invention, it is possible to effectively improve the player's ability by comparing the rating of the player with whether the player has solved the problem based on the rating of each of the problems, the tournament, and updating the database.

1 is a diagram showing an approximate structure and operation of a problem recommendation system according to an embodiment of the present invention.
FIG. 2 is a diagram showing a level corresponding to each problem, a tournament player, in a problem recommendation system according to an embodiment of the present invention. FIG.
Referring to FIG. 3, in a problem recommendation system according to an exemplary embodiment of the present invention, there is shown a probability that a player A is at an appropriate level for problems included in a specific competition based on a probability of winning.
4 is a flowchart illustrating an operation for recommending a problem based on player information input in a problem recommendation system according to an exemplary embodiment of the present invention.
FIG. 5 illustrates an example of a problem recommendation system according to an exemplary embodiment of the present invention. In FIG. 5, the proposed win probability and the comparison group win probability are sorted and the ratio of the problem solved by the player A It is a graph.
FIG. 6 is a graph showing the problems and the ratios of providing problems for each of the various players in a problem recommendation system according to an embodiment of the present invention, in a sequence close to 50%.

Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. The present invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, it is to be understood that the invention is not limited to the specific embodiments thereof, And equivalents and alternatives falling within the spirit and scope of the invention. In order to clearly illustrate the present invention in the drawings, parts not related to the description may be omitted, and the same reference numerals may be used for the same or similar components throughout the specification.

In various embodiments of the present invention, expressions such as 'or', 'at least one', etc. may denote one of the words listed together, or may represent a combination of two or more. For example, 'A or B', 'At least one of A and B' may include only one of A or B, and may include both A and B.

In various embodiments of the present invention, expressions such as 'first', 'second', 'first', 'second', etc. may describe various components, but they must mean the order, . For example, the first device and the second device are both devices and may represent different devices. Also, unless the elements of the configuration, function, operation, etc. of the first device are the same as or similar to the second device, the first device can be named as the second device, without departing from the scope of the various embodiments of the present invention, Similarly, the second device may also be termed the first device.

In the various embodiments of the present invention, when an element is referred to as being "connected" or "connected" to another element, the elements may be directly connected or connected, It should be understood that there may be one and the same time. On the other hand, if an element is referred to as being 'directly connected' or 'directly connected' to another element, it should be understood that no other element exists between the elements.

The terms used in various embodiments of the present invention are intended to illustrate a specific embodiment and are not to be construed as limiting the invention, for example, the singular forms "a," "an, ≪ / RTI >

In the following description, according to various embodiments of the present invention, it is possible to describe a system and an operation method thereof that provide a difficulty problem to an inputted (or designated) user (hereinafter referred to as a player).

The problem recommendation system may determine the rating of the tournament, including the rating (e.g., level) of the player and the rating of the problems contained in the database and / or at least one issue, It is possible to provide a problem corresponding to the determined player's rating.

The problem recommendation system can update (or update) the database using at least one regression analysis technique in determining the problem of the appropriate difficulty for the player based on the determined player's rating, the rating of the problems, have. Also, the problem recommendation system and its operation method according to various embodiments of the present invention may be interlocked with at least one other problem recommendation system, and may be included in other problem recommendation systems.

1 shows a schematic structure and operation of a problem recommendation system according to an embodiment of the present invention. Referring to FIG. 1, the problem recommendation system 100 may include at least one element of a processor 110, a player input 120, a memory 130, a communication unit 140, and a recommendation problem providing unit 150 have.

The processor 110 receives an instruction from the above-described other components (e.g., the player input unit 120, the memory 130, and the communication unit 150), decodes the received instruction, Or data processing can be executed.

For example, the processor 110 may extract a problem corresponding to a rating of a designated player based on the information input through the player input unit 120 and the database stored in the memory 130. [ In addition, the processor 110 may provide the extracted problem to the player through the recommendation problem providing unit 150. [

The processor 110 is configured to include at least a part of a rating determination unit 101 that determines the rating of the problems included in the memory 130 and a recommendation problem determination unit 103 that extracts a problem corresponding to the rating of the player . Here, it is described that the processor 110 includes the rating determiner 101 and the recommendation problem determiner 103, but this is for understanding the operation of the processor 110, and thus does not necessarily mean a physical configuration Do not. That is, the grading unit 101 and the recommendation problem determination unit 103 included in the processor 110 may be configured by a logic circuit corresponding to the operation.

The rating determining unit 101 may determine the rating (e.g., level) of the problems included in the memory 130, the ratings of the tournaments including the problems, the rating of the other players who have solved at least one problem, You can determine at least some of the ratings of the identified players.

According to one embodiment, the rating determiner 101 may determine the level of players solving the problem based on the ratings set in the issues. Also, players whose level has been determined can adjust the level of the problem they solved.

If the rating determination unit 101 determines that the player identified through the player input unit 120 is not a player stored in the memory 130, the rating determination unit 101 can provide a test problem composed of various difficulty levels.

For example, the grading unit 101 may randomly select and provide the problems included in the database regardless of the grades. According to another embodiment, the rating determiner 101 may classify various categories of difficulty based on the level of problems included in the database. For example, the rating determiner 101 can determine various difficulty problems as a test problem, including a problem configured in at least one difficulty level.

The rating determiner 101 may determine the player's level based on the pool input from the player. Here, the level of the player determined based on the test problem can be used as a temporary level to provide a recommendation problem.

The recommendation problem determination unit 103 determines whether or not the player has a probability that the pool probability of the problems included in the memory 130 based on the rating of the problems and the rating of the player, determined through the rating determination unit 101, , ≪ / RTI > or 0.5). The grading unit 101 can arrange the determined problems in the order close to the probability of 50% and determine the designated number of problems as the recommendation problem according to the descending order of the high probability.

The processor 110 may provide the player with the recommendation problems determined through the recommendation problem determination unit 103. [ The processor 110 may provide recommendation problems through at least one display, e.g., the recommendation problem offerer 150.

The recommendation problem providing unit 150 may be configured to include at least one display. The recommendation problem provider 150 can display various information to the player, for example, a recommendation problem determined through the recommendation problem determination unit 103. [ Also, the recommendation problem solving unit 150 may be configured as a touch screen for inputting a command by touching or proximity touching the input means on the display.

The communication unit 140 can connect communication between the problem recommendation system 100 and at least one external device. For example, the communication unit 140 can communicate with an external device by connecting to an existing network through wireless communication or wired communication, or by forming a new network.

Each of the aforementioned components of the problem recommendation system 100 according to various embodiments of the present invention may be composed of one or more components and the name of the component may be the type of the problem recommendation system 100 ≪ / RTI > The problem recommendation system 100 according to various embodiments of the present invention may be configured to include at least one of the above-described components, and some components may be omitted or further additional components may be included. In addition, some of the components of the problem recommendation system 100 according to various embodiments of the present invention may be combined into one entity so that functions of the corresponding components before being combined can be performed in the same manner have.

The problem recommendation system 100, which includes at least some of the components shown in FIG. 1 as described above, may provide recommendation problems to the player in various ways. According to one embodiment, the problem recommendation system 100 may use a player matching system to determine the problem of the appropriate degree of difficulty for the player.

For example, if the problem recommendation system 100 assumes that each of the problems contained in the memory 130 is a different player, then the player matching system may be used to determine a problem corresponding to the player's level. Here, the player matching system can determine the players' level based on the victory or defeat of the matched players.

On the other hand, the problem recommendation system 100 determines that the win has been won when the player solves the matched problem (e.g., when the correct answer is input), and when the matched problem can not be solved (e.g., , It is possible to apply the player matching system.

According to various embodiments of the present invention, the problem recommendation system 100 solves a problem or the like that is generated based on a k-factor raised in the Elo ranking technique (or the system thereof) . For example, with respect to the players A and B, it is possible to determine the ranking of the player A as R _A , the expected score of the player A as E _A , the rank of the player B as R _B , and the expected score of the player B as E _B.

Here, the expected score can be determined by whether the user (e.g., Player A) has solved the problem or not, for example, whether Player A has won Player B or not. In the following description, the problem recommendation target of the problem extraction system 100 can be defined as the player A. In this case, when the logistic curve is used, the expected score of the player A is expressed by the following equations (1)

(1)

(One)

.

Similarly, the expected score E _B of the player B is expressed by the following equations (2),

(2)

(2)

As shown in Fig.

Here, the player's ranking can be updated every time the game occurs. At this time, if Player A has an expected score of E _A and actually receives a score of S _A , the ranking of Player A (or the level of Player A)

(3)

(3)

Lt; / RTI >

Here, the k-factor can be determined as the maximum value of the adjustable score per game and is generally determined by a constant between the novice (e.g. 32) and the expert (e.g., 16) in determining the k- have. Therefore, it is difficult for the k-factor to determine an appropriate value among the novice or expert according to the player.

At this time, the k-factor may be varied according to the ranking of the player, and / or the number of times of confrontation. In the problem recommendation system 100, in determining the ranking of the players based on the k-factor, the problems are more frequent than the player, and in this case, for example, in the case of a high difficulty, As shown in FIG. On the other hand, the problem of low difficulty is placed in the lower part of the ranking because the number of competition is small because the number of recommendation is small.

In order to solve this problem, the problem recommendation system 100 can apply a Bayesian probability model. In the problem recommendation system 100, in determining the skill of each player and / or a participant, a gaussian distribution (Gaussian distribution) of players having an approximate player ability () and a degree of assurance () Distribution) can be used.

Here, [mu] can be determined as a class (e.g., level) designated for the player and / or the contestant. At this time, if? Is low, it can be determined that the reliability of? Increases. Each time a player plays a game, you can use the factor graph to update the player's σ and σ to apply the Beijian probability model.

In the Elo ranking system, only the player's rank is considered in ranking update, and the k-factor is seen in the constant feature where the past totals are not considered. On the other hand, the problem recommendation system 100 adopts the Bejingen probability model, meaning that a large number of past confrontation means that it is closer to the player's real skill, which makes σ smaller and σ is used in ranking update, The rank update can be performed.

According to various embodiments of the present invention, the problem recommendation system 100 determines whether the players stored in the database have solved or solved the problem in determining the parameters (parameters) of the system for recommending the problem to the player Can be used.

FIG. 2 shows, in a problem recommendation system 100 according to an embodiment of the present invention, each problem, a level corresponding to a tournament player.

Referring to FIG. 2, the problem recommendation system 100 may determine a level for each of at least one group of issues, competitions, and players stored in memory. For example, referring to the problem unit grade table 200, as shown in FIG. 2, a player can determine a level for a competition, including a problem, and at least one problem.

를 이고, 플레이어B의 기량이

이라면, 플레이어A가 플레이어B를 이길 확률은, 하기 식(4),Based on the determined grade and the designated variable, as shown in FIG. 2, the μ value and the sigma value of the problem can be trained with each player. For example, in problem recommendation system 100, player A's skill

And the skill of the player B is

, The probability that the player A will win the player B depends on the following equations (4) and

(4)

(4)

.

는 문제 추천 시스템(100)에서 기량과 성과의 관련이 높을수록 낮게 설정되는 상수일 수 있다. 또한,

함수는 정규 분포의 누적 분포함수로 나타낼 수 있다. 일 실시 예에 따르면, 트레이닝 시점의

, 초기

, 초기

을 설정할 수 있고, 전체 데이터의 80%에 해당하는 트레이닝 세트(Training set)으로 각 플레이어와 문제의

와

를 트레이닝할 수 있다.From here,

May be a constant that is set to be lower as the relation between performance and performance is higher in the problem recommendation system 100. [ Also,

The function can be expressed as a cumulative distribution function of the normal distribution. According to one embodiment,

, Early

, Early

Can be set, and a training set corresponding to 80% of the total data can be set for each player

Wow

Can be trained.

The problem recommendation system 100 may determine a probability (or a victory probability) to win against a pre-stored problem of a player of a training result (e.g. a test set, a test set) and a player (e.g., Player A) You can decide. The problem recommendation system 100 can provide the recommendation problem to the player A in the order close to the specified numerical value (for example, 50%, 0.5) based on the probability of winning determined based on the pre-stored problem. Here, if the value specified by the value specified as a value corresponding to the difficulty level of the recommendation problem (hereinafter referred to as the difficulty level of the problem recommendation) is 50% (or 0.5), the player A It is determined that the degree of difficulty is appropriate for the skill of the user. As described above, the difficulty level of the problem recommendation is 50%. However, it is obvious that various numerical values can be set to explain the understanding of recommending the problem of the difficulty appropriate to the skill of the player. For example, if the player A wants to provide a difficult problem more difficult than the player A, the difficulty level of the problem recommendation can be set to a value lower than 50%. If the player A wants to provide an easier problem than the player A, A higher value can be set.

와 실제 값

의 차이를 나타내는 것으로, 표준 편차(RMSE)는, 하기 식(5),The problem recommendation system 100 can determine the predictive performance of the problem recommendation based on whether or not the player A has solved the problem for the provided recommendation problem. According to one embodiment, the problem recommendation system 100 may determine the ratio at which player A solves the problem by comparing it with a reference value (e.g., 50%, 0.5) as it provides the recommendation problem to player A. For example, the predictive performance is such that Player A has a probability of

And actual value

, And the standard deviation (RMSE) is expressed by the following equations (5),

(5)Standard deviation (RMSE) =

(5)

. ≪ / RTI > The problem recommendation system 100 may update the database based on the determined information.

According to various embodiments, the problem recommendation system 100 may recommend a problem of the appropriate difficulty to the player in the tournament unit. The problem recommendation system 100 can determine a recommendation problem based on data of the player competition unit to recommend a problem of the competition unit.

For example, the problem recommendation system 100 can determine a problem corresponding to the difficulty level of the most difficult problem among the problems that the player A has solved in the tournament unit, as a problem suitable for the player. At this time, the level of the problems included in one tournament is determined based on at least one of the number of players who have solved each problem, the level of the tournament and the level of the player who solved the problem .

Referring to FIG. 3, it is possible to determine the probability that the problems (301) included in a specific tournament are appropriate level (adequate level) for the player A based on the probability that the player A will win. The problem recommendation system 100 may determine the problem of the degree of difficulty suitable for the player in the order of the problem or problem corresponding to the value of the highest probability based on the probability value that is the determined appropriate level.

FIG. 4 is a flowchart illustrating an operation for recommending a problem based on player information input by the problem recommendation system 100 according to an exemplary embodiment of the present invention.

Referring to step S401, the problem recommendation system 100 can input player information through the player input unit 120. [ In inputting the player information, the problem recommendation system 100 may receive at least some of various pieces of personal information such as name, age, sex, date of birth, and identification code of the player.

Referring to step S403, the problem recommendation system 100 can confirm whether or not the user is pre-stored (or registered) in the database based on the inputted player information. At this time, the problem recommendation system 100 can perform the step S409 if it confirms that the inputted player information is stored in advance, or can perform the step S405 if it confirms the player information that is not stored.

Referring to step S405, the problem recommendation system 100 may provide the player with a test question to determine the rating of the player. According to one embodiment, the problem recommendation system 100 may determine a test problem based on information such as age, gender, and the like of the player information entered. For example, the problem recommendation system 100 can identify other players that match the age and / or gender of the entered players from the database and determine test problems based on the level of other players identified.

Referring to step S407, the problem recommendation system 100 may determine the rating of the player based on the results of the solving of the test questions provided to the player. According to one embodiment, the problem recommendation system 100 can determine, through the rating determiner 101, a value obtained by averaging the level of the problem solved by the player as the player's level.

Referring to step S409, the problem recommendation system 100 may determine problems corresponding to the player's rating based on the players in the database and whether they have won or lost the issues. Here, whether or not the problem with the players is determined can be determined based on whether or not the players solved the problem.

For example, the problem recommendation system 100 may be configured to determine, based on the problems included in the memory 130, the number of times each problem has been solved, Can be confirmed. The problem recommendation system 100 can determine the problems that the players have solved as information related to victory or defeat, and can determine the grade of the problem based on the identified information.

The problem recommendation system 100 can use the grade of the problem and the rating of the player to determine the problems that the player has a probability of correcting the problem (the probability that the correct answer will be input) is close to 50% You can list up to the number of problems specified as a recommendation problem by listing them in close order of%.

According to various embodiments, the problem recommendation system can recommend a problem on a tournament basis, not limited to determining a recommendation problem on a problem basis, as described above, in accordance with an input through the player input unit 120. [

For example, the problem recommendation system 100 may determine a recommendation problem as a part of a problem that has been included in the tournament for at least one of the various tournaments included in the database.

According to one embodiment, the problem recommendation system 100 may be configured to provide a problem recommendation system 100 based on information such as the rating of the problems, the rating of the tournament including the problem and the rating of the player, You can determine the probability that the problems are at an appropriate level for the player. The problem recommendation system 100 can determine a specified number of problems as a recommendation problem by arranging the problems in a descending order of probability that they are appropriate levels for the problems.

Referring to step S411, the problem recommendation system 100 may provide the determined problems to the player. According to one embodiment, the problem recommendation system 100 may provide the player with a recommendation problem determined through the recommendation problematicizer 150. [

Referring to step S413, the problem recommendation system 100 may update the database and the player's rating based on the problem solved through the player input 120 for the provided recommendation problem.

According to one embodiment, the problem recommendation system 100 may determine whether or not an answer is correct for the entered information, and may determine whether the player has solved the problem, for example, whether the player has solved the problem Whether or not the answer is correct), and if the information of at least one of the rating of the problem solved by the player, the number of times the problem is solved, and the rating of the player is necessary based on the result of the solving, .

The problem recommendation system may terminate the embodiment of FIG. 4 by performing step S413.

According to various embodiments of the present invention, a simulation for performance analysis of the above-described embodiments can be performed. For example, the problem recommendation system 100 may perform a performance comparison, such as a prediction performance that determines the problem class of the appropriate degree of difficulty and the suitability of the recommendation problem, in determining a recommendation problem to the player.

In the simulation, the database uses datasets from 5,122 players in codeforces.com participating in 103 competitions. Dividing the data set arbitrarily, 80% is divided into training set and the remaining 20% is divided into test set.

,

,

로 결정할 수 있다.In the simulation, the problem recommendation system 100 defines each of the user and the problem as a player, and when the user (player A) solves the problem (player B), the winner of the player A, The player B can decide to win. In the proposed problem recommendation technique, in determining the probability that Player A will win Player B,

,

,

.

According to one embodiment, at least one other system for comparison with a problem recommendation system 100 (proposed problem recommendation system 100) according to various embodiments of the present invention may include, for example, matrix factorization (A comparative group problem recommendation system).

In general, a benchmark problem recommendation system can recommend a problem to a player based on a five-level matrix composition element that constitutes a player, a contest (eg, competition) and a contest. Here, the matrix elements of the 5 classes that make up the contest are used to determine the level of the toughest problem that the player has had in a particular contest, converted to a number (such as A-1, B-2, C-3, D- -5).

On the other hand, according to one embodiment of the present invention, the proposed problem recommendation system 100 determines whether the player has solved (for example, o) or not You can determine the level of a matrix element by converting it to a number (eg, o-1, x-0).

The range of these values can determine the active function of the matrix decomposition as a logical function and therefore the problem recommendation system 100 can use the SVDFeature to determine the following equation (6),

(6)

(6)

)을 결정할 수 있다.The probability of the player solving the problem by the comparative group problem recommendation system

Can be determined.

와

는 각각 플레이어A와 플레이어B(상술한 바와 같이 문제를 플레이어로 가정할 수 있다.)의 특성을 의미하고

와

는 각각 플레이어와 문제의 편향성을 의미하는 변수로 결정할 수 있다. 또한, 마지막으로

는 전체적인 경향성을 의미하는 변수이며, 시뮬레이션에서

는 60으로 설정했다.here

Wow

Respectively denote the characteristics of player A and player B (the problem can be assumed to be a player as described above)

Wow

Can be determined as a variable that indicates the bias of the problem with the player, respectively. Finally,

Is a variable indicating the overall tendency, and in the simulation

Lt; / RTI >

The proposed problem recommendation system 100 performs a problem-based simulation in which a player determined based on a recommendation problem (proposed problem) according to various embodiments of the present invention calculates the probability of solving a problem (Eg 1 or 0) of the probability that the player will solve the problem (comparison group win probability) and whether the player has solved the proposed problem, determined based on the problem recommendation technique using the matrix decomposition, Can be used to determine the standard deviation (RMSE).

According to the simulation described above, in the simulation of the problem unit, the prediction performance with respect to the proposed win probability and the actual value with respect to the comparison group win probability is obtained by a standard deviation (RMSE) of 0.281 and 0.33, respectively, as shown in Table 1 .

Suggested problem recommendation system Comparative group problem recommendation system Standard deviation (0-1) 0.281 0.33

To illustrate, a value closer to 0 indicates that there is no difference from the actual value, and it can be indicated that a recommendation of an appropriate level problem according to the level of the player. According to Table 1, the prediction performance of the proposed victory probability is 13% higher than that of the comparison group win probability.

In order to evaluate the suitability of the recommendation problem, FIG. 5 shows a case in which the user (player A) issues a question about the problem of the designated number (k) by sorting the proposed victory probability and the problem group This is a graph showing the ratio of the problem to be solved.

Referring to FIG. 5, in the recommendation of up to 5,000 problems, the value of the proposed winning probability area 501 is closer to 0.5 than the value of the comparison group winning probability area 503. This is because the proposed problem recommendation system 100 determines the problem of difficulty suitable for the level of the user A (player A) rather than the recommendation problem determined based on the comparative group technique up to 5,000 recommendation problems in determining the recommendation problem .

In addition, the proposed problem recommendation system 100 can recommend problems to various players. Therefore, it is necessary to extract and recommend a problem appropriate to each player's rating. FIG. 6 provides problems and shows the ratios for each of the various players according to the order in which the probability of problem solving is close to 50% (0.5).

According to one embodiment, when the problem recommendation system recommends a problem to a designated plurality of players, which is closest to 0.5 of the test set's problems, the proposed problem recommendation system exhibits a value of 0.002, Value, the proposed problem recommendation system 100 shows 30 times better performance than the comparative problem recommendation system. In the case of recommending more than 3 problems to the player, it can be seen that the difference between the comparative group problem recommendation system and the proposed problem recommendation system 100 is not much. On the other hand, when recommending a problem to a player, it is common to recommend a small number of problems, so that the performance of the proposed problem recommendation system 100 can be evaluated to be higher.

According to various embodiments, the proposed problem recommendation system 100 may be implemented in a simulation of a tournament unit, in which the level (recommendation problem and / or level of appropriate difficulty) of the recommendation problem according to various embodiments of the present invention, (RMSE) and accuracy based on the highest level of the problem that the player solved in the tournament (the problem and / or level of the top difficulty solved).

Here, the accuracy can be determined by averaging the values mapped to 0 in one case when the level of the problem of the appropriate level of difficulty is the same as the level of the problem in the best level, at a predicted hit ratio of the difficulty level suitable for the player.

In the proposed problem recommendation system 100, for each of the proposed problem recommendation system 100 and the comparative group problem recommendation system, the level of the determined appropriate difficulty level and the level of the resolved best difficulty level , Standard deviation (RMSE), and accuracy are shown in Table (2).

Comparative group problem recommendation system Suggested problem recommendation system Standard deviation (0-5) 1.02678 0.67157 Accuracy (0-1) 0.5449677 0.7594

Referring to Table 2, it can be seen that the results of the proposed problem recommendation system 100 are improved by 31% in terms of standard deviation (RMSE) and by 79% in terms of accuracy in comparison with the results of the comparative group problem recommendation system.

According to various embodiments, at least some of the devices and methods according to the various embodiments described in the claims of the present invention and / or the specification are in the form of hardware, software, firmware, or a combination of two or more of hardware, For example, modules, units). A module may be a minimum unit or a portion thereof that performs various embodiments of the present invention as a minimum unit or a part of an integrally constructed component. The module may be implemented mechanically or electronically. When implemented in software, a computer-readable storage medium (or computer-readable storage medium) for storing one or more programs (or programming modules, applications) may be provided. For example, the software may be embodied in instructions stored on a computer-readable storage medium in the form of a programming module. The one or more programs may include instructions to cause the device to perform the methods according to the embodiments of the invention and / or the claims of the present invention. The instructions, when executed by one or more processors (e.g., the processor 220), may cause the one or more processors to perform functions corresponding to the instructions. The computer readable storage medium may be, for example, the memory 230). At least some of the programming modules may be implemented (e.g., executed) by, for example, the processor 220. At least some of the programming modules may include, for example, modules, programs, routines, sets of instructions or processes, etc. to perform one or more functions.

The computer-readable recording medium includes a magnetic medium such as a hard disk, a floppy disk and a magnetic tape, an optical recording medium such as a CD-ROM (Compact Disc Read Only Memory), a DVD (Digital Versatile Disc) A magneto-optical medium such as a floppy disk and a program command such as a ROM (Read Only Memory), a RAM (Random Access Memory), a flash memory, (EEPROM), a magnetic disc storage device or other type of optical storage device, or any other form of optical storage device, A magnetic cassette may be included. Or a combination of some or all of these. In addition, a plurality of constituent memories may be included.

In addition, the device may be accessed via a communication network comprised of a communication network such as the Internet, an Intranet, a Local Area Network (LAN), a Wide Area Network (WLAN), or a Storage Area Network (SAN) and can be stored in an attachable storage device that can access the storage device. Such a storage device may be connected to the device via an external port. Also, a separate storage device on the communication network may be connected to the portable device. The hardware devices described above may be configured to operate as one or more software modules to perform operations on the various embodiments of the present invention, and vice versa.

Modules or programming modules according to various embodiments of the present invention may include at least one or more of the elements described above, some of which may be omitted, or may further include other additional elements. Operations performed by modules, programming modules, or other components in accordance with various embodiments of the invention may be performed in a sequential, parallel, iterative, or heuristic manner. Also, some operations may be performed in a different order, omitted, or other operations may be added.

The embodiments of the present invention disclosed in the present specification and drawings are merely illustrative examples of the present invention and are not intended to limit the scope of the present invention in order to facilitate understanding of the present invention. Accordingly, the scope of the present invention should be construed as being included in the scope of the present invention, all changes or modifications derived from the technical idea of the present invention.

100: problem recommendation system 110: processor
120: player input unit 130: memory
140: communication unit 150: recommended problem provisioning
101: rating decision unit 103: recommended problem decision unit

Claims (14)

Confirming the degree of assurance of the rating based on the rating of the inputted player and the problems that have been solved and the total score;
Calculating a probability that the player wins the problems based on the player and the degree of assurance of each of the problems and the rating of each of the problems;
Comparing the calculated victory probability with a predetermined matching recommendation value to sort the problems for the player in an order close to the matching recommendation value; And
And providing the player with a predetermined number of questions according to the sorting order.
The method according to claim 1,
The rating of the player is determined based on the input information of the player, and when the rating of the player is not specified,
Based on a test problem provided based on at least one of the problems, the at least one corresponding to two or more difficulty levels.
The method according to claim 1,
Wherein the rating of the problems comprises at least one of a level of the problems, a level of the tournament that includes the problems, and a problem matching method based on the player matching technique.
The method of claim 3,
Determining a probability that the problems are at an appropriate level for the player based on the rating of the problems and the rating of the player; And
And providing the player with a designated number in the order of higher probability of the appropriate level.
The method according to claim 1,
Whether or not the above-
When the player inputs a correct answer to the question, it is determined that the player has won,
Wherein the player is determined to have been defeated if the player has entered an incorrect answer to the question.
The method according to claim 1,
Wherein the step of determining a probability that the player wins the problems based on the rating of the player and the rating of each of the problems and the rating,
Wherein the degree of certainty of the player and the problems and the degree of the problems is determined by applying the following equation.

Equation

(Where P _AB is the probability that A will beat B,

Is a constant determined based on rank and performance,

Is the cumulative distribution function of the normal distribution,

Is the rating of the player and the problems that have been resolved,

The degree of confidence in the rating)
The method according to claim 1,
Updating the win or loss of the player with respect to the provided problem to the database; And
And adjusting the rating of the player based on the win / lose of the player.
A memory containing a database of problems;
A player input unit for receiving information from the player;
A recommendation problem providing unit for providing problems to the player; And
A processor coupled to the memory, the player input, and the recommendation problem offerer,
The processor comprising:
Checking the degree of confidence of the player based on the rating of the player who has entered the player through the player input unit and the problems of the problems included in the database, And comparing the calculated victory probability with a predetermined matching recommendation value to sort the problems with respect to the player in the order close to the matching recommendation value And processes the player to provide a predetermined number of problems to the player in accordance with the sorting order through the problem providing unit.
9. The method of claim 8,
The processor determines the rating of the player based on the input information of the player, and if the rating of the player is not specified,
Wherein the rating of the player is based on a test problem provided comprising at least one corresponding to at least two difficulties of the problems.
9. The method of claim 8,
Wherein the processor includes at least one of a level of the problems and a level of a tournament that includes the problems in the grade of the problems.
11. The method of claim 10,
The processor is configured to determine a probability that the problems are at an appropriate level for the player based on the rating of the problems and the rating of the player and to provide the player with a designated number in an order that is more likely to be the appropriate level , Problem recommendation system based on player matching technique.
9. The method of claim 8,
Wherein the processor determines that the player has won when the player has entered a correct answer to the question and determines that the player has been defeated if the player has entered a wrong answer to the question, system.
9. The method of claim 8,
Wherein the processor applies a degree of assurance of the player and the class of the problems and the degree of confidence to the following equation to determine the probability that the player wins the problems.

Equation

(Where P _AB is the probability that A will beat B,

Is a constant determined based on rank and performance,

Is the cumulative distribution function of the normal distribution,

Is the rating of the player and the problems that have been resolved,

The degree of confidence in the rating)
9. The method of claim 8,
Wherein the processor is configured to update the winner and the loser of the player against the provided problem to the database and to adjust the player's rating based on win or loss of the player.

Images