KR101096431B1 - Method and system of classifying books - Google Patents

Abstract

PURPOSE: A method and a system for classifying a book are provided to increase the efficiency of a category classification by automatically classifying a category of a book based on a category learning database. CONSTITUTION: A category learning database(150) receives and stores book data including category of a classified book and a book name from a book information database(200). The category learning database includes distribution information between words configuring book name of the classified book and category. An extracting unit(130) extracts the combination of words which is included in the book name of the classified target book from the book name of a classified target book. A category estimator(140) estimates the category of the classified target book by using the category learning database and the combination of the words.

Description

Book classification method and book classification system {Method and System of Classifying Books}

The present invention relates to a book classification method and a book classification system, and more particularly, to a book classification method and system capable of automatically classifying a book category from a signature of a book.

Recently, each library that owns a large number of books, such as a university library or a public library, recommends a book for purchase, or obtains information on a new book from a company that distributes the book, or a publisher. It is true.

In purchasing a book like this, it is necessary to purchase a book suitable for each library. Books listed in each library are classified by category, and the library can consider the weight of each category of books to be purchased as an important factor considering the current weight of each category. If so, it is necessary to properly classify the categories of books to be purchased.

However, it is very cumbersome to classify categories manually by hand for books to be purchased. In addition, even if the same book can be classified into different categories, it is unlikely that it would be inconsistent with the actual classification of the library if it is classified by a third party (e.g., a book supplier) who is not in charge of the classification work of the library. There is also a problem.

In addition, if you want to classify directly at the library to be purchased, the books to be purchased must include a number of books that do not lead to actual purchase, and the classification task of classifying the categories for books that are not actually purchased is very troublesome. There is.

In addition, when a physical book, which is actually purchased, arrives at the library, the book is classified according to the classification criteria or classification behavior of the library when the book is received. However, this sorting work requires a lot of time and effort, but there is no way to alleviate such sorting work.

On the other hand, since the perception of the problems and problems of the prior art is not obvious to those skilled in the art of the present invention, judging the progress of the present invention compared to the prior art based on this perception Make no sense.

SUMMARY OF THE INVENTION The present invention was created to solve the above problems, and an object of the present invention is to provide a method and system capable of automatically classifying book categories.

Another object of the present invention is to provide a method and system for automatically classifying book categories according to the classification criteria or classification behavior of a specific library.

In addition, another object of the present invention is to provide a method and system that can reduce the category classification work of books in the purchase process and the receipt of the book.

A book classification method according to an aspect of the present invention is a book classification method performed by a computer or a server in association with a book information database 10 storing book information, and a category is already classified from the book information database 10. Receiving book data including at least a category and signature of a mood book, which is a book, and using the book data, wherein the category learning database is a distribution between each word and category constituting at least the signature of the mood book Includes information-a first step of constructing; A step 2-1 of receiving a signature of a book to be classified as a book to classify a category and extracting a set of words from the signature of the book to be classified; And a step 2-2 of estimating a category of the book to be classified using the set of words extracted in the step 2-1 and the category learning database.

A book classification system according to an aspect of the present invention is a book classification system for automatically classifying books to be classified in association with a book information database 200 that stores book information. A category learning database 150 constructed by receiving book data including at least a category and a signature of a mood book, which are already classified books, and including distribution information between each word and category constituting at least the signature of the mood book ; Extraction means (130) for extracting a set of words included in the signature of the book to be classified from the signature of the book to be classified which is the book to be classified; And a category estimating means (140) for estimating the category of the book to be classified using the set of words extracted by the extracting means (130) and the category learning database.

According to an aspect of the present invention, since the categories of books can be automatically classified, there is an effect that can significantly increase the efficiency of category classification.

In addition, according to an aspect of the present invention, it is possible to automatically classify the categories of books in accordance with the classification criteria or classification behavior of a particular library, and accordingly there is an effect that can be categorized according to the situation of the particular library.

In addition, according to an aspect of the present invention, there is an effect that can reduce the category classification work of the book in the purchase process and the receipt of the book.

1 is a block diagram illustrating a book classification system 100 and a book information database 200 according to an embodiment of the present invention.
FIG. 2 is a flow chart illustrating a book classification method according to an embodiment of the present invention. FIG. 2 (A) illustrates the construction of a category learning database, and FIG. 2 (B) illustrates a method of classifying categories of books to be classified. Detailed process is shown.
3 is a flowchart illustrating a process of building the category learning database 150 in detail.
4 is a diagram illustrating an example of the appearance count of each word in each category.
5 is a diagram illustrating an example of an appearance frequency rate calculated based on the appearance count of FIG. 4.
FIG. 6 is a flowchart illustrating a specific process of estimating a category of a book to be classified using a set of extracted words and a category learning database.

Hereinafter, with reference to the drawings will be described a preferred embodiment of the present invention. In addition, the Example described below does not unduly limit the content of this invention described in the claim, and the whole structure demonstrated by this embodiment is not necessarily required as a solution of this invention.

1. Book Classification System

1 is a block diagram illustrating a book classification system 100 and a book information database 200 according to an embodiment of the present invention.

The book information database 200 stores book information and is, for example, a database that is constructed for the management of books in a university library or a public library. The book information may be a signature, a category, a date of purchase, a publisher, an author, a summary, a book review, and the like for the book. In particular, according to an embodiment of the present invention, the book information includes at least information about a book's signature and category. The 'category' of a book may be according to various classification methods such as Korean Decimal Classification (KDC), Dewey Decimal Classification (DDC), and International Decimal Classification (UDC).

The book classification system 100 according to an embodiment of the present invention performs a function of automatically classifying books to be classified in association with the book information database 200, and may be implemented as a general computer or a server.

The input / output means 110 is responsible for functions such as data transmission and reception with the book information database 200, input / output with an administrator or user, input / output with another computer or server (not shown), and the like. In particular, the input / output means 110 receives book data from the book information database 10, and the book data includes at least the signature and category of the book.

Hereinafter, like a book registered in the book information database 100 described above, a book whose category has been classified and the category is defined is referred to as a 'classified book'. On the other hand, books that have not yet been categorized are referred to as "books to classify". The book to be classified is a book to be categorized using the book classification method or book classification system of the present invention. In particular, the books to be classified may be books directly related to an institution such as a college library or a public library using the book information database 200. For example, books that are likely to be purchased by these institutions, books that are expected to be received by these institutions, books purchased by these institutions, and the like.

The learning means 120 performs a function of constructing a category learning database 150 using book data input from the book information database 100. Specifically, the learning means 120 includes at least a category of mood books that are books that have already been categorized. Distribution information between each word and category constituting the signature of the mood book is constructed from the book data including the signature.

The category learning database 150 performs a function of storing distribution information between each word and category constructed by the learning means 120. Specifically, the category learning database 150 is constructed from book data including at least a category and a signature of a mood book, which is a book whose category is already classified, and distribution information between each word and category constituting the signature of the mood book. Save it.

The extracting unit 130 performs a function of extracting a set of words included in the signature of the book to be classified from the signature of the book to be classified, which is the book to classify the category. Performs a function to extract a set of words included in a signature.

The category estimating means 140 performs a function of estimating the category of the book to be classified using the set of words extracted by the extracting means 130 and the category learning database 150.

The storage means 160 stores the results classified by the category estimating means 140 and performs a function of storing data and temporary data necessary for performing the book classification system 100.

Specific functions and additional functions of the book classification system 100 may be better understood through the description of the book classification method, and the above description is not described for all the functions of the book classification system 100. Care must be taken.

2. How to sort your books

FIG. 2 is a flow chart illustrating a book classification method according to an embodiment of the present invention. FIG. 2 (A) illustrates the construction of a category learning database, and FIG. 2 (B) illustrates a method of classifying categories of books to be classified. Detailed process is shown.

First, as shown in FIG. 2A, a process of building a category learning database (S100) must be preceded.

Book classification method according to an embodiment of the present invention is carried out in conjunction with the book information database 10 that stores the book information, the book information database 10 stores the book information, for example, university library or It is a database being built for the management of books in public libraries. The book information may be a signature, a category, a date of purchase, a publisher, an author, a summary, a book review, and the like for the book. In particular, according to an embodiment of the present invention, the book information includes at least information about a book's signature and category.

The process of constructing the category learning database 150 is performed by using the book data, which receives book data including at least a category and a signature of a mood book, which is a book whose category is already classified, from the book information database 10. The category learning database 150 includes at least distribution information between each word and category constituting the signature of the mood book. In the above description, 'build' does not mean that a person performs, but refers to a book classification system 100 according to an embodiment of the present invention, specifically, the learning means 120.

3 is a flowchart illustrating a process of building the category learning database 150 in detail.

First, the book classification system 100, specifically, for example, the input / output means 110 receives book data of a mood book from the book information database 200 (S110).

In addition, the book classification system 100, specifically, for example, the extraction unit 130 extracts a set of words from the signature of the mood book (S120). When the set of words is extracted from the signature of the mood book in step S120, the set of extracted words may be a set of self-supporting morphemes, real morphemes, or lexical morphemes.

A morpheme is the smallest unit containing meaning in a language and is a language unit that loses its meaning when further analyzed. Like phonemes, morphemes are abstract entities and can be realized in various forms in speech.

The morphemes are divided into free morphemes, bound morphemes, and full morphemes and empty morphemes according to the loss of meaning. Real morphemes that represent lexical meanings, such as specific objects, actions, and states, are called lexical morphemes. do.

A lexical morpheme is a morpheme that has a lexical meaning and refers to an object, state, or action. Generally, nouns, verbs, adjectives, and adverbs belong to this category. For example, in the words "there is a lot of good information in the library," "library," "good," "information," and "many," are lexical morphemes.

Freestanding morphemes are morphemes that can be used alone without other morphemes. In Korean, nouns generally belong to this. In the example above, “library” and “information” are the equivalent.

In addition, the book classification system 100, specifically, for example, the learning means 120 performs a process of increasing the appearance count for each word belonging to the set of extracted words in the corresponding category of the mood book by one. .

The above-described steps S120 and S130 may be repeatedly performed on the plurality of mood books, for example, until there are no more mood books to learn (see step S140). In addition, when the book data of the mood book is input one by one in step S110, the process returns to step S110.

In addition, step S100 may further include step S150. In step S150, the book classification system 100 calculates the frequency of appearance of each word included in the signatures of all mood books (S150). Specifically, for each category of mood books, the frequency of occurrence of each word included in the signature of all mood books of the category is calculated, where the frequency of occurrence of the word is a certain category. Refers to the frequency of occurrence of the word in the signature of the classification book.

□ Frequency of occurrence of word A in category C

= (Count of occurrences of word A in the signature of a book in category C) / (count of appearances of all words in the signature of a book in category C)

Fig. 4 is a diagram showing an example of appearance counts for each word in each category.

For example, there are 10 books in category 1, and the signature of these books is 5 for word A, 2 for word B, 2 for word C, 2 for word D, 1 for word E, and 1 for word. F appears once, and the appearance counts are 5, 2, 2, 2, 1, and 1, respectively.

5 is a diagram illustrating an example of an appearance frequency rate calculated based on the appearance count of FIG. 4.

For example, in category 1, the frequency of word A is 5/13, the frequency of word B is 2/13, the frequency of word C is 2/13, and the frequency of word D is 2/13, the appearance frequency of word E is 1/13, and the appearance frequency of word F is 1/13.

In order to obtain the appearance frequency rate, when the appearance frequency rate becomes zero (ZERO), a problem that it may be difficult to estimate the category to be described below may occur. Therefore, a slight correction of the appearance frequency rate obtained above is performed. You may.

In the above formula for calculating the frequency of appearance, a method of additionally adding 1 to the numerator and additionally adding the appearance count of the word +1 to the denominator is used. The above-mentioned occurrence rate calculation formula may be modified as follows.

□ Frequency of occurrence of word A in category C (variation)

= (Count of occurrences of word A in the signature of a book in category C + 1) / (count of appearances of all words in a signature of a book in category C + word A in the signature of a book in category C Appearance count of + 1)

The category learning database 150 stores distribution information between each word and category constituting the signature of the mood book, and the distribution information may be the appearance count information or the appearance frequency information.

2, the detailed process of the method for classifying the category of the book to be classified shown in FIG. 2 (B) will be described.

First, the book classification system 100 receives a signature of a book to be classified (S200). For example, the input and output means 110 of the book classification system 100 may be input from another computer or server (not shown), or may be directly input from an administrator or user of the book classification system 100.

In addition, the book classification system 100, specifically, for example, the extraction unit 130 receives a signature of a book to be classified as a book to classify a category and extracts a set of words from the signature of the book to be classified (S300). The process of extracting the set of words from the signature is similar to the process of extracting the set of words from the signature of the mood book in step S120.

When extracting a set of words from a signature of a book to be classified, which is a book to classify, and extracting a word from the signature of the book to be classified, the set of words to be extracted may be a self-supporting morpheme, a real morpheme or a lexical morpheme.

In addition, the book classification system 100, for example, the category estimation unit 140 estimates the category of the book to be classified using the extracted word set and the category learning database 150 (S400).

FIG. 6 is a diagram illustrating a specific process of estimating a category of a book to be classified using a set of extracted words and a category learning database 150.

First, the book classification system 100 calculates a proportional probability for each category, which is a value proportional to the probability that the book to be classified is a category for each of all categories that can be classified (S410). The proportional probability by category can be obtained by multiplying the occurrence frequency rate and the category occupancy rate for all words belonging to the extracted word set. Here, the appearance frequency rate refers to the "appearance frequency rate" obtained in step S150, and uses the one already established in the category learning database 150. On the other hand, the category learning database 150 may store and store only the appearance count of each word in each category, and calculate the above-mentioned appearance frequency rate in real time in the process of estimating the category from the actual signature later.

The 'category share' is the ratio of the mood books of the category to the total number of mood books. For example, in the example shown in FIG. 4, the category occupancy of category 1 is 0.1 as 10 / (10 + 20 + 25 + 30 + 15).

For example, assuming that the signature of the classification book includes the words A, B, A, F, and E, the proportional probability for each category may be calculated as follows.

□ Proportional Probability of Category 1

= 10/100 * (5 + 1) / (13 + 6) * (2 + 1) / (13 + 3) * (5 + 1) / (13 + 6) * (1 + 1) / (13+ 2) * (1 + 1) / (13 + 2) = 0.000033241

□ proportional probability of category 2

= 20/100 * (2 + 1) / (16 + 3) * (3 + 1) / (16 + 4) * (2 + 1) / (16 + 3) * (2 + 1) / (16+ 3) * (1 + 1) / (16 + 2) = 0.000017495

□ Proportional Probability of Category 3 = 25/100 * (1 + 1) / (21 + 2) * (2 + 1) / (21 + 3) * (1 + 1) / (21 + 2) * (6+ 1) / (21 + 7) * (5 + 1) / (21 + 6) = 0.000013127

□ Proportional Probability of Category 4 = 30/100 * (2 + 1) / (13 + 3) * (2 + 1) / (13 + 3) * (2 + 1) / (13 + 3) * (3+ 1) / (13 + 4) * (1 + 1) / (13 + 2) = 0.000062040

□ Proportional Probability of Category 5 = 15/100 * (4 + 1) / (18 + 5) * (1 + 1) / (18 + 2) * (4 + 1) / (18 + 5) * (1+ 1) / (18 + 2) * (6 + 1) / (18 + 7) = 0.000019849

The book classification system 100 estimates the category of the proportional probability having the largest value among the proportional probabilities for each category obtained above as the category of the book to be classified (S420). In the above calculation example, since the proportional probability of category 4 has the largest value, books containing the words A, B, A, F, and E in the signature are classified as category 4. And as another embodiment, the method of recommending a category in the order of high proportional probability can also be used. In the example above, the ranking of category 4 and category 1 may be recommended.

The estimated category is once stored in the storage means 160, and may be transmitted to another computer or server through the input / output means 110 or used by the book classification system 100 by itself.

3. Mathematical theory of the present invention

Hereinafter, a mathematical theory of a book classification method according to an embodiment of the present invention will be described.

In probability theory, P (A | B) P (B) = P (B | A) P (A), which is modified, P (A | B) = P (B | A) P (A) / P Becomes (B).

Here, P (A | B) is a conditional probability that event A will occur while event B occurs, and P (B | A) is a conditional probability that event B will occur while event A occurs. Also, P (A) is the probability that event A will occur and P (B) is the probability that event B will occur.

Applying this to the probability that the book with the signature D = (t ₁ , ..., t _n ) is category C, Equation 1 can be obtained.

Where (t ₁ , ..., t _n ) is the set of words included in the signature, P (C | D) is the conditional probability of category C when signature D, and P (t ₁ , ..., t _n | C) is the conditional probability that signature D will occur when category C, P (C) is the probability that category C will occur, and P (t ₁ , ..., t _n ) is the probability that signature D will occur.

Meanwhile, the following Equation 2 can be obtained based on Equation 1 above.

는 가능한 카테고리들 중에서 그 후단에서 계산되는 값이 최고가 되는 카테고리를 의미한다. 따라서 위 수학식 2는

가 최대가 되는 카테고리는

가 최대가 되는 카테고리와 동일하다는 의미를 가진다. 여기서

는 서명 D일 때 카테고리 C일 확률이며, P(t_i|C)는 카테고리 C일 대 단어 t_i의 확률로서, 실제 계산에 있어서는 전술한 단어의 '출현 빈도율'로서 대표될 수 있다.At this time,

Denotes a category in which the value calculated at a later stage among the possible categories becomes the highest. Therefore, Equation 2 above

Is the largest category

Has the same meaning as the maximum category. here

Is the probability of category C when signature D, and P (t _i | C) is the probability of category C versus word t _i , and can be represented as the 'frequency of occurrence' of the words described above in actual calculation.

P (C) is the probability that category C will occur and can be represented as the aforementioned 'category share' in the actual calculation. In Equation 1, P (t ₁ , ..., t _n ) on the right side is deleted from Equation 2 because it is a category-independent factor.

Therefore, according to equation (2),

가 최대로 되는 카테고리를 구하면 결국 서명 D가 속할 가능성이 가장 높은 카테고리를 구할 수 있게 되는 것이다.

If we find the category of which is the maximum, we can find the category most likely to belong to the signature D.

는 실제 계산에 있어서 전술한 카테고리별 '비례 확률'로 대표될 수 있으며, 이 비례 확률이 최대가 되는 카테고리가 서명 D가 속할 가능성이 가장 높은 카테고리가 되는 것이다.
Meanwhile,

Can be represented by the above-mentioned 'proportional probability' for each category in the actual calculation, and the category in which the proportional probability is maximum becomes the category most likely to belong to the signature D.

Hereinafter, the effects of the invention according to various aspects of the present invention will be described in detail.

According to an aspect of the present invention, since the categories of books can be automatically classified, there is an effect that can significantly increase the efficiency of category classification.

On the other hand, even in the case of using the same classification system for each library due to the nature of the classification work, the same books are often classified into different categories in performing the task of classifying the actual books. Libraries may establish their own classification criteria or categorize books with their own classification behavior in order to maintain consistency in categorization.

According to an aspect of the present invention, since a category learning database learned in association with a book information database of books of a specific library is constructed and classified using the same, the book category is automatically matched to the classification criteria or classification behavior of the specific library. There is an effect that can be classified and thus the category can be tailored to the specific library.

In addition, according to an aspect of the present invention, it is possible to automatically recommend the category of the book by the system, accordingly there is an effect that the person in charge of the classification task to automatically provide information that can be referred to before performing the correct classification by manual . According to an aspect of the present invention, there is an effect that can reduce the category classification work of the book in the purchase process and the receipt of the book.

100: book classification system 110: input and output means
120: learning means 130: extraction means
140: Category Estimation Means 150: Category Learning Database
160: storage means 200: book information database

Claims (14)

As a book classification method performed on a computer or server in association with a book information database (10) storing book information,
The computer or the server receives from the book information database 10 book data containing at least a category and signature of a mood book, a category of which is already classified, and using the book data, a category learning database-a category The learning database includes at least a distribution information between each word and category constituting the signature of the mood book;
A step 2-1 of the computer or the server receiving a signature of a book to be classified as a book to classify a category and extracting a set of words from the signature of the book to be classified;
A step 2-2 of the computer or the server estimating a category of the book to be classified using the set of words extracted in the step 2-1 and the category learning database;
The first step is,
For each category of mood books, the frequency of occurrence of each word in the signature of all mood books in the category, where the frequency of occurrence of the word is that word in the signature of all mood books in that category when in a particular category. Refers to the frequency of appearance-to obtain the first 1-3 steps;
The second step 2-2,
Calculating a proportional probability for each category, which is a value proportional to the probability that the book to be classified is a category for each of all categories that can be classified;
And a step 2-2-2 of estimating the category of the proportional probability having the largest value among the proportional probabilities of the categories obtained in the step 2-2-1 as the category of the book to be classified.
In calculating the proportional probability for each category in step 2-2-1,
The proportional probability of each category is the frequency of occurrence and category occupancy for all words belonging to the set of words extracted in step 2-1. The percentage of occupancy is obtained by multiplying the book classification method.
The method according to claim 1,
The first step is,
Extracting a set of words from the signature of the mood book;
In the corresponding category of the mood book, Steps 1 and 2 for increasing the appearance count for each word belonging to the set of words extracted in the step 1-1;
The method of claim 1, wherein the steps 1-1 and 1-2 are repeated for a plurality of mood-like books.
delete The method according to claim 1,
In order to avoid the problem that the appearance frequency becomes 0 when calculating the appearance frequency of each word in the first to third steps,
A formula for calculating the frequency of appearance, wherein the numerator of the formula is (the count of the occurrences of the corresponding word appearing in the signature of the book included in a particular category) and the denominator of the formula is the sum of Appearance count)
And additionally adding 1 to the numerator and (addition count +1 of the words appearing in the signature of books included in a particular category) to the denominator.
The method according to claim 2,
When extracting a set of words from the signature of the mood book in step 1-1, the extracted set of words is a collection of independent morphemes, real morphemes or lexical morphemes.
Claim 1
When extracting a set of words from the signature of the book to be classified, which is a book to classify categories in step 2-1, and extracting a set of words from the signature of the book to be classified, the set of words to be extracted is an independent morpheme, a real morpheme or a lexical morpheme. Book classification method, characterized in that the set of.
delete delete A book classification system for automatically classifying books to be classified in association with a book information database (200) storing book information,
The book information database 10 is constructed by receiving book data including at least a category and a signature of a mood book, which is a book whose category has already been classified, and at least distribution between each word and category constituting the signature of the mood book. A category learning database 150 containing information;
Extraction means (130) for extracting a set of words included in the signature of the book to be classified from the signature of the book to be classified which is the book to be classified;
And category estimating means (140) for estimating the category of the book subject to classification using the set of words extracted by the extracting means (130) and the category learning database.
Distribution information of the category learning database 150,
For each category of mood books, the frequency of occurrence of each word included in the signature of all mood books in that category, where the frequency of occurrence of the word corresponds to the signature of all mood books in that category for a particular category. Refers to the frequency of occurrence of the word-,
The proportional probability for each category is the frequency of occurrence and category occupancy for all words belonging to the set of words extracted by the extraction means 130, where the category occupancy is the mood book of the category in the number of all mood books. The ratio of occupied is obtained by multiplying and estimating the category of the proportional probability by category having the largest value among the obtained proportional probability by category as the category of the book to be classified.
delete delete delete The method according to claim 1,
Book information stored in the book information database 10 relates to books already owned by a particular library,
The book to be classified is a book classification method, characterized in that the book is likely to be purchased by the particular library, books that are to be received in the particular library or books purchased by the particular library.
As a book classification method performed on a computer or server in association with a book information database (10) storing book information,
The computer or the server receives from the book information database 10 book data containing at least a category and signature of a mood book, a category of which is already classified, and using the book data, a category learning database-a category The learning database includes at least a distribution information between each word and category constituting the signature of the mood book;
A step 2-1 of the computer or the server receiving a signature of a book to be classified as a book to classify a category and extracting a set of words from the signature of the book to be classified;
A step 2-2 of the computer or the server estimating a category of the book to be classified using the set of words extracted in the step 2-1 and the category learning database;
The first step is,
For each category of mood books, the frequency of occurrence of each word in the signature of all mood books in the category, where the frequency of occurrence of the word is that word in the signature of all mood books in that category when in a particular category. Refers to the frequency of appearance-to obtain the first 1-3 steps;
The second step 2-2,
Calculating a proportional probability for each category, which is a value proportional to the probability that the book to be classified is a category for each of all categories that can be classified;
From the categories of proportional probability having the largest value among the proportional probabilities obtained in step 2-2-1, two or more categories are sequentially estimated as categories of the book to be classified and used to recommend two or more categories by rank. Including; 2-2-2 step of allowing;
In calculating the proportional probability for each category in step 2-2-1,
The proportional probability of each category is the frequency of occurrence and category occupancy for all words belonging to the set of words extracted in step 2-1. The percentage of occupancy is obtained by multiplying the book classification method.

Images