Artificial Intellegence

Project: Data grouping with java

Create an application visualizing groups creation according to a hierarchical algorithm. Every object is described by two numeric attributes. Both the input objects and output of the algorithm

$($

grouped objects

$)$

$$svvhould be presented graphically. Grouping should be made according to the shortest distance between two objects Project: Data grouping

Create an application visualizing groups creation according to a hierarchical algorithm.

Every object is described by two numeric attributes. Both the input objects and output of the

algorithm $($grouped objects$)$ should be presented graphically. Grouping should be made according to

the shortest distance between two objects. The distance between objects $p_i$ and $p_j$ is calculated

according to the relation:

$d_{ij}=\sqrt[\phantom{2}]{{\displaystyle \underset{a=1}{\overset{N_a}{}}}(p_{ia}-p_{ja}{)}^2}$

where:

$N_a-$ number of attributes,

$,p_{ia},p_{ja}-$ value of attribute a for objects i and $j.$

Data is loaded from files in the RSES format $(*.$ tab$).$ If there are more than two attributes, the

application should allow the selection of appropriate attributes. The final number of groups should

be given by the user.

The program should calculate statistics related to the correctness of created groups.

Additional functionality of the application $($for each realized $-$ half grade more$)$:

grouping objects with more than $2$ attributes,

ability to perform clustering with weights $($see Appendix$),$

text information about steps performed by the algorithm,

stepping visualization of joints of different groups.Artificial Intelligence

Laboratory, full$-$time studies, year $2023/2024$

Project: Data grouping

Create an application visualizing groups creation according to a hierarchical algorithm.

Every object is described by two numeric attributes. Both the input objects and output of the

algorithm $($grouped objects$)$ should be presented graphically. Grouping should be made according to

the shortest distance between two objects. The distance between objects $p_i$ and $p_j$ is calculated

according to the relation:

$d_{ij}=\sqrt[\phantom{2}]{{\displaystyle \underset{a=1}{\overset{N_e}{}}}(p_{ia}-p_{ja}{)}^2}$

where:

$N_a-$ number of attributes,

$,p_{ie},p_{ja}-$ value of attribute a for objects i and $j.$

Data is loaded from files in the RSES format $(*.$ tab$).$ If there are more than two attributes, the

application should allow the selection of appropriate attributes. The final number of groups should

be given by the user.

The program should calculate statistics related to the correctness of created groups.

Additional functionality of the application $($for each realized $-$ half grade more$)$:

grouping objects with more than $2$ attributes,

ability to perform clustering with weights $($see Appendix$),$

text information about steps performed by the algorithm,

stepping visualization of joints of different groups.

Appendix

The distance with weights between objects $p_i$ and $p_j$ is calculated according to the relation:

$d_{ij}=\sqrt[\phantom{2}]{{\displaystyle \underset{a=1}{\overset{N_j}{}}}{W}_a*(p_{ia}-p_{ja}{)}^2},$

where:

$N_a-$ number of attributes,

$p_{ia,}{p}_j-$ value of attribute a for objects i and $j,$

$W_a-$weight for attribute $a.$

Weight $W_a$ is calculated with:

$W_a=\frac{d_a}{\text{m}\text{a}\text{x}(d_a,d_b,d_c,\text{d}\text{o}\text{t}\text{s})},$

where:

$a,b,c-$ attributes set.

Distance $d_a$ is calculated as:

$d_a=ma{x}_a-mi{n}_a,$

where:

$ma{x}_a-$ maximum value for the attribute a among all the objects,

$mi{n}_a-$ minimum value for the attribute a among all the objects.