1. Introduction
In classical set theory, an element belongs or does not belong to a set. Such a binary distinction is inadequate to deal with many uncertain real-life situations. Different generalized set concepts are defined to handle such situations, among them fuzzy sets  [
1], soft sets  [
2], intuitionistic fuzzy sets  [
3], and rough sets  [
4,
5].
Rough set theory addresses classification problems when the exact membership of elements is uncertain. The technique uses equivalence relations that group elements in equivalence classes, which are also called indistinguishable classes. Then, the target set is approximated by creating two basic approaches:
	  
- Lower approximation: Set of equivalence classes covered by the target set. Elements belonging to such equivalence classes are assumed to be members of the set with certainty; there is no doubt about their membership. 
- Upper approximation: This is a set consisting of equivalence classes that have a nonempty intersection with the target set. The elements in these classes are highly probable members of the set, but there is uncertainty about their exact membership. 
The difference between the lower and upper approximations is called the boundary region and reflects the unreliability in classification. If the boundary region is not an empty set, the target set is a rough set. Additionally, an “accuracy value”, with values ranging from 0 to 1, is used to measure the effectiveness of classification within a data set  [
6].
However, defining an equivalence relation for every uncertain scenario is not always feasible. For this reason, a binary relation was used instead of an equivalence relation, and the 
j-neighborhood concepts were defined instead of an equivalence class  [
7,
8].
Since the accuracy value in rough set theory represents a measure of success, achieving an accuracy value closer to “1” has become a primary objective for researchers. As a result, numerous generalizations of 
j-neighborhood have been developed to improve classification performance and address various complexities in data analysis. Some of these are 
-neighborhoods  [
9], subset neighborhoods  [
10], maximal neighborhoods  [
11,
12], and 
-neighborhoods  [
13]. The new defined neighborhood concepts are used to reduce uncertainty. For example, the authors developed previous models inspired by 
-neighborhoods using the concept of ideals in  [
13]. They demonstrated that their proposed models expand the amount of confirmed information and reduce ambiguity, leading to more accurate decisions.
Topology is one of the tools used in studying rough set models due to the similarity between topological operators and the approximation operators of these models. Many studies have explored these concepts and the relationships between them in both domains. Additionally, some topological concepts, such as generalizations of open sets, have been applied to the study of these models, as demonstrated in studies  [
14,
15,
16,
17,
18]. Furthermore, topological generalizations have been employed to address practical problems in these models. For instance, the authors of  [
19,
20] utilized the concepts of supratopology and infratopology to describe rough set models.
Rough set theory is particularly useful in fields such as data mining  [
21], pattern recognition  [
22], artificial intelligence  [
23], and decision making to prevent the spread of some epidemics, such as COVID-19  [
24,
25,
26,
27], where clear decisions are difficult due to the complexity or fuzziness of the data.
Although many types of neighborhoods have been defined as generalizations of rough sets in the literature, there has not been much work specifically addressing situations where elements are not directly related to each other. This research introduces a new concept in rough set theory called the j-transitive neighborhood (abbreviated as j-TN) to achieve more effective results in such cases. In previous studies, the hybridization of rough neighborhoods focused on the same types of neighborhoods. This study, however, links different types of neighborhoods, a novel approach that allows for the definition of new rough set models. These models will provide tools to address certain practical problems that appear in medical science and social media
The sections of this study are as follows. The first section is the introduction. The second section of this study is dedicated to providing the necessary foundational knowledge on rough set theory. In the third section, j-TN is defined, its properties are examined, and examples are provided to support this. Furthermore, various analyses are performed on the sets of transitive elements that ensure that two elements are j-TN, and certain results are obtained based on the properties of the binary relation. In the fourth section, the concepts of lower and upper approximations are defined using the j-TN concept. Consequently, the accuracy values of a set are derived, and these values are compared in tables with the approximation and accuracy values obtained from j-neighborhood and j-TN. In the fifth section, the lower and upper approximation concepts are redefined on the topological structure obtained through j-TN. Furthermore, comparisons are made with previously defined topological structures in the literature, evaluated through examples. In the sixth section, the concept j -TN is applied to identify indirect contacts and classify disease risks to take the necessary precautions to break the transmission chain in an infectious disease. Finally, the seventh section is dedicated to the conclusion and future work.
  3. j-Transitive Neighborhood
This section presents the concept of j-transitive neighborhood (abbreviated j-TN).
Definition 8.  Let  The j-transitive neighborhoods of a point  are defined as follows:
.
.
.
.
.
.
.
.
 Proposition 2.  If  is a symmetric relation, then  Proof.  This result follows from the fact that if  is symmetric, then     □
 Remark 2.  If multiple binary relations are defined on a set X, we refer to their j-transitive neighborhoods by the relation’s name. For instance, let  and  be two relations on X. Then,  and  are the r-transitive neighborhood sets of the point x with respect to the relations  and , respectively.
 Example 1.  Let  and . The j-neighborhoods and j-transitive neighborhoods of the points of X are given in Table 1.  The following example shows that the concepts j-neighborhood and j-transitive neighborhood are not comparable under the subset relation.
Example 2.  Let  and  Then, , but  Besides, , but  Let  Then,  and  But  and 
 Proposition 3.  Let  and  Then, the following hold:
- (i) 
-  and . 
- (ii) 
- , . 
 Proof.  (i) Straightforward.
		   
- (ii)
-  For , the results are obtained in a similar manner. 
□
 Lemma 1.  Let  and  be two relations on X such that . Then,for .  Proof.  Assume 
 and let 
. This implies the existence of 
 such that 
. Consequently, 
, leading to 
. Similarly, it can be observed that 
. Moreover, since 
 and 
, it follows that
        
        and
        
□
 Example 3.  Let  andThe j-transitive neighborhoods for  based on  and  are given in Table 2 and Table 3, respectively. As evident from the provided example, the assertion (3) presented in Lemma 1 holds. However, their conversations are typically not valid.  Result 1.  Example 3 illustrates that the converse of (3) in Lemma 1 does not hold, and it is invalid for   Definition 9.  Let  and  be a mapping such that  for all . Then  is called a j-transitive neighborhood space, or briefly, a j-TN space.
 Proposition 4.  Let  be a j-TN space. Then,  and  for all 
 Proof.  Let . This implies . Consequently,  and . Thus, there exist  such that . As a result,  and . Therefore, it follows that  and .    □
 Result 2.  Let  be a j-TN space. Then, the following hold:
- (i) 
- . 
- (ii) 
- . 
 Proposition 5.  Let  be a j-TN space and . Then, the following hypotheses are held:
- (i) 
- If  is reflexive, then  for each . 
- (ii) 
- If  is symmetric and , then  and  for . 
- (iii) 
- If  is transitive and , then  
- (iv) 
- If  is transitive,  and , then  
- (v) 
- If  is antisymmetric and , for all , then  is reflexive. 
- (vi) 
- If  is an equivalence, then , for all . 
 Proof.  (i) It is straightforward.
		   
- (ii)
- Since , . Given the symmetry of , it follows that . Consequently,  and . 
- (iii)
- Let . Then , such that  Since  is transitive,  Therefore,  
- (iv)
- From (iii), . Since , we obtain  
- (v)
- Assume that  is not reflexive. Then,  such that  Since  then there exists  such that  However, because  is antisymmetric, it follows that  which contradicts the assumption that  Therefore,  is reflexive. 
- (vi)
- Let  be an equivalence relation. From (i), it follows that , for . Furthermore, since , for , this implies , for  
□
 Following, the correlation between the j-transitive neighborhoods based on the relation  and its inverse, denoted as , is demonstrated.
Proposition 6.  Let  be a j-TN space. Then
- (i) 
- if  is symmetric, then  
- (ii) 
 Proof.  (i) The desired result is obtained directly, as  when  is symmetric.
		   
□
 The following definition is introduced to characterize elements that facilitate the transition between two given elements in a j-TN space.
Definition 10.  Let  be a binary relation on X. If , then z is called a transition element from x to y. The set of all transition elements from x to y is denoted by . It follows that  and 
 Proposition 7.  Let  and  Then  Proof.  Straightforward.    □
 Proposition 8.  Let  be a j-TN space.
- (i) 
- If , then  and  
- (ii) 
- If  and  is symmetric, then . 
- (iii) 
- If  is symmetric and , then  
- (iv) 
- If  is reflexive and , then  and . 
- (v) 
- If , then  and  
- (vi) 
- If , then . 
- (vii) 
- If  then  
- (viii) 
- If  is antisymmetric and , then  
 Proof.  (i) It is clear from Definition 8 and Definition 10.
		   
- (ii)
- Let  and  be symmetric. Then  Therefore, . 
- (iii)
- Since  and  is symmetric, then . Therefore,  
- (iv)
- Since  is reflexive and , then  Therefore,  and  
- (v)
- Let , then . Therefore,  and . Hence,  and  
- (vi)
- Let . Then  Thus, . 
- (vii)
- Let  Then  Therefore,  This implies that  
- (viii)
- Let . Then,  Since  is antisymmetric, it follows that  
□
   4. Rough Approximation Based on j-Transitive Neighborhood
This section gives the concepts of j-transitive lower and j-transitive upper approximations utilizing transitive neighborhoods.
From this point onward, let  be a j-TN space.
Definition 11.  The j-transitive lower and j-transitive upper approximations with respect to  of a subset  are defined as, respectively,for all   Proposition 9.  Let  be a j-TN space, and . The following conditions hold for all :
- (i) 
- . 
- (ii) 
-  and . 
- (iii) 
- If , then  and . 
- (iv) 
- . 
- (v) 
- . 
- (vi) 
- . 
- (vii) 
- . 
 Proof.  Statements (i) and (ii) are obvious.
        
- (iii)
- Let . This implies that  or  for some  such that . Since , we have  or , which implies . - Let  Since , we have  for all . Therefore,  
- (iv)
- Let . Then  or . This implies that  or , for some . Thus, we have . As a result, . - It can be derived from (iii) that  Let  Then,  or  for all  such that . This implies that  or  such that  or [ or  such that ]. Thus,  or . Consequently, it follows that  
- (v)
- As established in (iii), we have  and  
- (vi)
- (vii)
- From statement (i), we derive that . For the sake of contradiction, assume that . Then, there exists some  such that . This implies that  for all  that satisfy . However, since , we have , which is a contradiction. Therefore, it must be the case that . - Similarly, the statement  can be derived. 
□
 The presented example illustrates that the equality of statement (ii) in Proposition 9 does not hold.
Example 4.  Let  and . Then, , for all .
 The following examples demonstrate that statements (iv) and (v) in Proposition 9 do not have equality.
Example 5.  Let  and  be subsets of the j-TN space  where , . Then, , while . Moreover, , but 
 Example 6.  Let  and  be subsets of the j-TN space  where  and  Then  However, .
 Example 7.  Table 4 and Table 5 give all j-transitive neighborhoods of subsets of X in Example 1.  Definition 12.  The -boundary, -positive, and -negative regions of a subset  are defined as, respectively,  Definition 13.  The  accuracy of  is defined aswhere  If , then A is called  exact set; otherwise, it is called  rough.  Definition 14.  Table 6 and Table 7 give the j-accuracy and -accuracy of all subsets of X defined in Example 1 for each .    5. Topological Structure and Approximations Derived from j-Transitive Neighborhood
This section defines the topology , referred to as -topology, on a j-TN space using j-transitive neighborhood relationships and subsequently investigates various associated characteristics. In addition, approximations within -topological spaces are established and analyzed.
Theorem 1.  Let  be a j-TN space. Then, the setis a topology on X, for all .  is called as j-TN topological space.  Proof.  It is evident that  Let  and . For all , it follows that  and  Hence, , implying  Now, consider  for  where  is an index set. Let  There exists  such that  Consequently, , for all . This implies that  establishes that 
□
 Example 8.  In the set X provided in Example 1, the topologies  and  with respect to j-neighborhoods and j-transitive neighborhoods become as follows:
.
 The following theorem is derived from properties (i) and (ii) presented in Result 2.
Theorem 2.   be a j-TN topological space. Then, the following hold:
- (i) 
- . 
- (ii) 
- . 
- (iii) 
- . 
- (iv) 
- . 
 Example 8 illustrates that the converse of relations given in Theorem 2 is invalid.
Theorem 3.  Let  be a j-TN topological space. Then  and 
 Proof.  Suppose . This implies that for all , . Consequently, we have . Therefore, for all , it follows that  and . Hence, . In contrast, we can demonstrate that , which completes the proof.
Similarly, it can be shown that     □
 Theorem 4.   be a j-TN topological space. Then,  is the dual topology of 
 Proof.  Let  and  Let us assume that  Then, there exists  Therefore, by Proposition 3,  Since  and , then  This means , which is a contradiction. Consequently, .    □
 Theorem 5.  Let  be a j-topological space and  be a j-TN topological space. Then, we have , for .
 Proof.  () Let , and let . Consider . It follows that , which implies the existence of . Consequently,  and . Given that  and , we can deduce that , thus establishing that . Furthermore, since  and , we can conclude that . As a result, , which establishes .
() Analogously, we have .
() Let , and . We observe that . Consequently,  and . For any , it follows that . This implies either  or . In turn, this implies the existence of  or . Thus,  with  or  with . In both cases, we have  and  or  with . Therefore, we conclude that , establishing that . This completes the proof.
□
 The following examples illustrate that the converse of Theorem 5 is not generally true and is not valid for 
Example 9.  Consider the set  with the relation . In this context, we observe that , while  and .
 Example 10.  Let  and . Then, , , while , , and 
 Example 11.  Let  and . Then, , and . Therefore, , but 
 Remark 3.  From Example 8, it can be inferred that , and  for .
 Result 3.  The following inclusions hold:
- (i) 
- . 
- (ii) 
- . 
 Below, the approximation concepts relevant to the topology derived from j-transitive neighborhoods are defined.
Definition 15.  The j-lower and j-upper approximations of a subset A of a j-TN topological space  are defined as  Proposition 10.  Let  be a j-TN topological space and . Then, the subsequent properties hold:
- (i) 
-  and . 
- (ii) 
- . 
- (iii) 
- If , then  and . 
- (iv) 
-  and . 
- (v) 
-  and . 
- (vi) 
-  and . 
 Proof.  The proofs of properties , , and  are derived from Definition 15.
        
- (iv)
- From , we have . To show the reverse inclusion, let . Then, there exist  such that  and . Hence,  Since , it follows that  - Similarly, we can prove that . 
- (v)
- Since  and , we have . Similarly, . Conversely, for any , there exists  such that  and . This implies that  and . Thus,  Therefore, . 
- (vi)
- Since , it follows that . To show the converse, assume  Then,  for some  Given that , we have . Thus, . - Similarly, we can prove that . 
□
 Example 12.  Table 8 and Table 9 give the j-approximations of all subsets of the topological space  given in Example 1.  Definition 16.  Let  be a j-TN-topological space. The regions and measures of subsets are derived using  and , analogous to the corresponding Definition 13.
 Example 13.  The j-accuracy of all subsets of the topological space  and the topological space , as presented in Example 1, are provided in Table 10 and Table 11, respectively.    6. A Medical Application
Both direct and indirect contact play a major role in the spreading of infectious diseases throughout a community. Direct contact refers to the physical contact between an infected individual and another person that may lead to the fast-spreading of diseases. For instance, physical contact such as handshaking or kissing is effective in transferring viruses and bacteria. Indirect contact is the transmission of disease through an intermediate agent. One may not have contact with the infected person directly but nevertheless be exposed to a risk of the disease through the third party that has been in contact with an infected person. For instance, in a case where individual A is infected and has direct contact with individual B, and later on, individual C comes into contact with individual B, individual C is considered to have indirect contact with the disease of individual A. Differentiating the modes of transmission, especially direct and indirect contact, is thus paramount in infectious disease control.
The j-neighborhood concept has been applied to various decision-making tools in studies that aim to identify and isolate contacts to control infectious disease spread and break the transmission chain  [
29]. These studies have mainly focused on direct contacts. However, determining indirect contacts becomes especially essential in those illnesses with a long incubation period where indirect contact remains asymptomatic for a longer period of time, so preventive measures may be taken more effectively.
A scenario is considered below where an epidemic is present. The concept of j-transitive neighborhoods (j-TN) is adapted to identify indirect contacts and implement necessary measures to mitigate the spread of the epidemic.
Consider the set X, which comprises individuals working in an institution and their families (or people they are in close contact with). When an infectious disease starts spreading among employees, those who test positive for the disease are immediately placed in quarantine. Let  represent the set of infected people. To mitigate the risk of further transmission, people who have had direct contact with infected patients are also placed under observation. Let  denote the set of these direct contacts.
Now, our objective is to investigate the risk of infection among those who are not infected or who have not had direct contact with infected individuals. This subset, represented as , includes employees and their families who have not directly interacted with infected patients.
We define a relation on X such that “ is in direct contact with y and ”. Since  is symmetric, we have only two types of j-transitive neighborhoods by Proposition 2. Therefore, the set of indirect contacts of x is .
To classify each  according to their risk of infection, we define the following risk levels (these numbers can be changed with an expert opinion) based on the number of contacts.
Low Risk: The number of contacts is between 0 and 3 (inclusive).
Moderate Risk: The number of contacts is between 4 and 7 (inclusive).
High Risk: The number of contacts exceeds 7.
To better understand this scenario, we examine the following example. Let 
 and 
. In 
Table 12, the symbol “✓” is used to indicate direct contact within the set 
X.
According to the 
Table 12, the set of direct contacts is 
, and the set of indirect contacts is 
. The 
l-transitive neighborhoods of the individuals in 
T are given below:
.
.
.
.
.
.
.
- Since  T-  is the set of indirect contacts of  E- , no individual in  T-  has an  l- -neighborhood in the set  E- . Thus, the classifications derived using the concepts of  l- -transitive neighborhood and  l- -neighborhood, and the given risk levels above are presented in  Table 13- . 
Table 13 reveals that individuals who should be categorized as high-risk are instead classified as low-risk when the 
j-neighborhood concept is applied.
   7. Conclusions and Future Work
In order to cope with the cases where it is not clear whether an element belongs to a set, equivalence classes are examined and three different cases are determined: the element is not a member of the set, it is definitely a member, or it is probably a member. By introducing the concepts of right and left neighborhoods, the requirement for an equivalence relation is eliminated, allowing for a reduction in uncertainty in more complex or ambiguous situations. For this purpose, the neighborhood concept has been generalized in many studies, and the rough set concept has been reconsidered from different perspectives  [
30,
31]. In almost all neighborhood concepts defined previously in the literature, other elements directly related to an element have been examined. The 
j-transitive neighborhood (
j-TN) concept defined in this study focuses on resolving uncertainties where indirect relationships are involved. The main results obtained within the scope of the study are summarized below:
* As demonstrated in Example 1, the concepts of j-neighborhood and j-transitive neighborhood are not directly related.
* Theorem 9 shows that, unlike certain generalized approximations present in the literature  [
9,
11], the lower–upper approximation concepts derived using 
j-TN satisfy the properties of the Pawlak approximation, as described in Proposition 1, except the equality in (P5).
* 
Table 6 and 
Table 7 show that a clear comparison cannot be made between the 
j-accuracy and 
 accuracy values. However, in terms of the topological accuracy values examined in 
Table 11, a higher accuracy value is obtained in the case of 
.
* The concept of 
j-transitive neighborhoods (j-TN) was examined in the context of preventing the spread of infectious diseases, where indirect contact plays a critical role. Indirect contacts were grouped according to their risk of transmitting the disease. The classification revealed that individuals categorized as high-risk under the 
j-transitive neighborhood were classified as low-risk when evaluated using the 
j-neighborhood (see 
Table 13). This observation indicates that the defined neighborhood concept contributes to a more accurate categorization of at-risk individuals.
In a future study, more comprehensive analyses of the approximations obtained with j-TN will be carried out, and these approximations will be compared with the approximation concepts defined previously in the literature. In addition, some topological concepts such as generalized open sets, supratopology, and infratopology are planned to be adapted to the j-TN concept. Since the hybridization of rough neighborhoods with ideal improves the properties of approximation operators and increases the accuracy value, one can examine the current concepts from this perspective.