Non-hierarchical Method Research Articles

ANALYSIS OF SEGMENTATION AND TARGETING OF JEANS IN JAKARTA

Jeans are a significant part of the global apparel industry, with a vast market. This research aims to analyze how the jeans segmentation is formed and who exactly is the target market. The population in this study were all groups who wear jeans in Jakarta. The technique used in determining the sample was purposive sampling with a total of 161 people. The type of research used in this research is descriptive research using the hierarchical procedure method to search for segmentation. This analysis aims to see whether there are significant differences between each cluster or group formed. Then, in searching for target markets, this research uses a non-hierarchical procedure method. From the results of the analysis, 2 clusters were formed where the 2nd cluster was selected as the target market. It is hoped that the results of this analysis can provide input to jeans business actors to be able to attract more potential consumers and accurately. This study employs sentiment analysis to delve into the preferences and opinions of jeans users in Jakarta, offering fresh insights for businesses to develop more effective marketing strategies and understand market trends, consumer preferences, and growth potential.

ANALYSIS OF RAINFALL IN INDONESIA USING A TIME SERIES-BASED CLUSTERING APPROACH

Indonesia has a tropical climate and has two seasons: dry and rainy. Prolonged drought can cause drought disasters, and rain can cause floods and landslides. According to information from the Meteorology, Climatology, and Geophysics Agency (BMKG), natural disasters such as floods and landslides due to heavy rains have been a severe problem in Indonesia for the past five years. Different regional characteristics can affect the intensity of rain that falls in every province in Indonesia. It can be grouped to determine which provinces have similar characteristics to natural disasters due to rainfall. Later, it can provide information to the government and the public so that they are more aware of natural disasters. So, it is necessary to research and classify provinces in Indonesia for rainfall with cluster analysis. The data used is secondary rainfall data taken from the official BMKG website. Cluster analysis of rainfall in 34 provinces in Indonesia used hierarchical and non-hierarchical methods in this study. The approach that is used in this research limits our clustering of the data. Further research with a machine learning approach is recommended. For the clustering method, the agglomerative hierarchical method includes single, average, and complete linkage. The non-hierarchical method includes k-medoids and fuzzy c-means. The cluster analysis results show that the dynamic time warping (DTW) distance measurement method with the average linkage method has the most optimal cluster results with a silhouette coefficient value of 0.813.

Classification of Healthy Family Indicators in Indonesia Based on a K-means Cluster Analysis.

Health development is a key element of national development. The goal of improving health development at the societal level will be readily achieved if it is directed from the smallest social unit, namely the family. This was the goal of the Healthy Indonesia Program with a Family Approach. The objective of the study was to analyze variables of family health indicators across all provinces in Indonesia to identify provincial disparities based on the status of healthy families. This study examined secondary data for 2021 from the Indonesia Health Profile, provided by the Ministry of Health of the Republic of Indonesia, and from the 2021 welfare statistics by Statistics Indonesia (BPS). From these sources, we identified 10 variables for analysis using the k-means method, a non-hierarchical method of cluster analysis. The results of the cluster analysis of healthy family indicators yielded 5 clusters. In general, cluster 1 (Papua and West Papua Provinces) had the lowest average achievements for healthy family indicators, while cluster 5 (Jakarta Province) had the highest indicator scores. In Indonesia, disparities in healthy family indicators persist. Nutrition, maternal health, and child health are among the indicators that require government attention.

Analisis pengelompokan jumlah tanaman kehutanan yang diusahakan menurut jenis tanaman di Indonesia

One of the problems of archipelagic countries is the lack of maximum utilization of natural resources which has resulted in some areas being left behind. Indonesia is one of those who experience the impact of the lack of utilization of natural resources in the forestry sector. The non-optimal use of forests for planting forestry plants has made most Indonesians use their land as artificial forests, namely to plant forestry plants. Cluster analysis in this case seeks to classify provinces in Indonesia based on the type of forestry plants cultivated. The method used is hierarchical and non-hierarchical. The hierarchical method uses single linkage and complete linkage methods while the non-hierarchical method uses the K-mean method. By making comparisons between methods, the results obtained are that the single linkage method with 8 clusters is the best method for grouping provinces in Indonesia according to the types of plants cultivated. Of the 34 provinces in Indonesia, cluster 1 consists of 27 provinces, while clusters 2 to 8 each consist of only 1 province.
 Keywords: Cluster analysis, single linkage, complete linkage, K-mean, forest plantsMSC2020: 62H30

Classification of Datasets Based on Combination Algorithm of Clustering and Neural Network

How to get the needed information from the data accurately and effectively for analysis is a hot research topic nowadays. Accurate classification of data is the basis for smooth data analysis. In order to classify data effectively, scholars have proposed some classification algorithms, and the most frequently mentioned one is k-means algorithm. However, in previous studies, scholars have directly determined the number of classes to be classified for the data set to be classified. Therefore, in this paper, a combinatorial algorithm is proposed to improve the classification of data with unknown group classes. The clustering algorithm and neural network are also combined to improve classification accuracy. The main elements of the algorithm proposed in this paper are as follows. First, one-third of a set of unknown group class data is selected as the sample data. In order to accurately assess the characteristics of a set of anonymous group class data, it is important first to choose a sample of the data. A sample of one-third of the total data set should be sufficient to provide a reliable representation of the entire population. This sample should be selected at random in order to ensure that the results of the assessment are as accurate as possible. The number of classes is determined by using hierarchical clustering method on the sample data. Then, the non-hierarchical clustering k-means method is used to classify the sample data. Finally, the classification results are trained as the training items of the neural network, and then the model generated after the training is used to classify the overall data. This paper selects three datasets with different kinds, different numbers of variables, and different amounts of data for the experiments and analysis. This paper presents a comprehensive analysis of three distinct datasets. Each dataset has its own unique characteristics, such as its type, the number of variables, and the amount of data contained therein. By leveraging the properties of these datasets, the experiments, and analysis conducted in this paper will provide valuable insights into the data structures and trends contained within. Furthermore, the results from this analysis will serve as a foundation for further research and experimentation. The experimental results show that the combination of clustering algorithm and neural network algorithm will help to improve the accuracy of data classification and identification effectively. This research provides a new way to accurately and effectively perform data classification. Keywords: hierarchical clustering; k-means; neural network; data classification.

NON HIERARCHICAL K-MEANS ANALYSIS TO CLUSTERING PRIORITY DISTRIBUTION OF FUEL SUBSIDIES IN INDONESIA

The growth rate of inflation in Indonesia continues to increase from day to day. The inflation rate in Indonesia reached 1.17% in September 2022 which is the highest inflation rate in the last seven years. One of the causes of high inflation is caused by the increasing demand for motor vehicle fuel. Therefore, there is a need for appropriate action from the government in determining related policies. K-Means multivariate cluster analysis is a non-hierarchical cluster method that is popularly used, one of which is used in Machine Learning algorithms, especially Unsupervised Learning. The purpose of this research is to clustering that are priority distribution of subsidies in Indonesia based on the characteristics formed. The data in this study consist of the percentage of poverty, the percentage of total transportation, the percentage of transportation use, and the percentage of area. Data were analyzed using multivariate cluster analysis with the K-Means method. Based on the research results, information was obtained that the data fulfilled a representative sample with value of KMO >50%. In addition, there are 4 optimal clusters which are the results of the calculation of the Elbow and Silhoutte methods, so 4 provincial clusters are formed with their respective characteristics. Cluster 1 is a province that is highly prioritized to receive fuel subsidies, Cluster 2 is a province that is not highly prioritized for fuel subsidies, Cluster 3 is a province that is prioritized to receive fuel subsidies, and Cluster 4 is a province that is not prioritized to receive fuel subsidies.

Hierarchical classification for acoustic scenes using deep learning

Acoustic Scene Classification (ASC) aims to obtain the sound environment by analyzing audio signals. Due to the low complexity and acquisition cost of audio signals, ASC has enormous potential in various applications, such as audio-based surveillance, smart cities/homes, and robotics. Recently, various methods have been proposed for ASC and achieved good performance. However, when they are used to address complex ASC problems, most of them suffer from the low-performance problem. In this paper, we propose to use hierarchical classification methods to replace the conventional flat approach in ASC applications, which utilizes the class hierarchy to optimize classification performance. In particular, we investigate the ASC problem under the framework of hierarchical classification. Firstly, to improve classification performance, three hierarchical classification methods introducing the class hierarchy of acoustic scenes are proposed for ASC. Moreover, to fully utilize the class hierarchy, a hybrid hierarchical classification method, and an optimal late fusion-based hierarchical method are proposed, which are based on the flexibility and simplification of hierarchical classification. The experiments demonstrate the efficacy of hierarchical ASC systems for performance improvement, and the best system achieves an accuracy of 78.86% on the DCASE 2020 Task1A dataset, resulting in accuracy gains of 24.76% and 8.52% absolute over the DCASE 2020 Task 1A baseline and the conventional non-hierarchical method, respectively.

K-means Nonhierarchical Cluster and Dbscan Outlier Detection In the Grouping of Stock Issuers

Group analysis aims to group objects based on similar characteristics so that they are in one group homogeneous and between groups heterogeneous. Study this aim group issuer share in Indonesia based on volatility, liquidity, and market capital. This study uses the non-hierarchical K-Means Clustering method, because the number of samples is big and the number of groups are known. The K-Means Clustering grouping method produces as many as 6 groups with different characteristics. 2. Group 1 consists of stock issuers with quite high volatility and liquidity. The characteristic of group 2 is that it consists of stock issuers with the lowest volatility. Big Capital is the nickname for group 3 because it has market capital or the asset value is very large among all groups and the volatility is very small, and liquid. In Group 4, stock issuers have the highest volatility and the lowest liquidity. Results of profile interpretation in group 5, issuer’s stocks have the highest liquidity and market capital is quite low. Share issuers in group 6 have the volatility highest. Group 3 is recommended as an option for investing. Because, having market capital or large asset values, liquid, and volatility is low enough to minimize risk. The originality of this research is that there is no combination of methods between grouping in fields with k-means clustering and detection of sales with DBSCAN, especially in the field of issuer share in Indonesia.

Multivariate analysis of post-transition OECD countries in the context of inequality measures

The aim of this paper is to classify the post-transition OECD countries according to the Gini coefficient for income inequality, the S80/S20 ratio, the income share of the bottom 40% of the population, educational attainment – tertiary education, and labor force participation rate using factor and cluster analyses. Factor analysis resulted in two extracted factors, and factor scores were calculated. Hierarchical and non-hierarchical cluster analysis was performed on factor scores to classify eight posttransition OECD countries and three candidate countries. The research question of this paper is to investigate whether there are similarities/differences between existing members and candidates for membership in the OECD, among the selected post-transition countries of Europe, in the context of income inequality. Based on the dendrogram obtained by the hierarchical Ward’s method, a three cluster solution was selected. The non-hierarchical k-means method for the three-cluster solution clustered Croatia with Bulgaria and Romania. These three countries are OECD candidate countries. Our findings confirm that the three candidate countries remain behind because of historical reasons and the non-implementation of structural reforms.

850-P: The Efficacy of Imeglimin in Diabetes Subgroups Stratified by Data-Driven Cluster Analysis—A Post Hoc Analysis of Imeglimin Clinical Trial Data

Objectives: We performed a post-hoc analysis of clinical trial data with the goal of investigating whether diabetic subgroups identified by data-driven clustering respond differently to imeglimin. Methods: Data from randomized, double-blind clinical trials of imeglimin as monotherapy or add-on to insulin therapy were included in the analysis. For monotherapy, 1) duration of type 2 diabetes, 2) baseline BMI, 3) baseline HbA1c, and 4) baseline HOMA-β, and for add-on to insulin therapy, the first three and 4) insulin total daily dose were applied as coordinates to form clusters by the non-hierarchical k-means method. The efficacy of imeglimin were examined for each cluster. Results: We identified four clusters of patients with diabetes, which had significantly different patient characteristics. Cluster 1 was a group with lower values of all four indices compared to the total patient population before cluster segregation, cluster 2 was a group with longer duration of diabetes, cluster 3 was a group with higher baseline BMI and higher HOMA-β (for monotherapy) or higher insulin total daily dose (for add-on to insulin therapy), and cluster 4 was a group with higher baseline HbA1c. The improvement in HbA1c with imeglimin differed significantly among the four clusters. The mean change differences versus placebo for imeglimin in monotherapy ranged from -0.64% to -1.27%, and add-on to insulin therapy ranged from -0.31% to -0.82%. In all clusters except for cluster 3 in the add-on to insulin therapy, imeglimin showed statistically significant HbA1c improvement compared to placebo. Conclusion: Data-driven clustering of patients with type 2 diabetes allowed us to identify subgroups that respond differently to imeglimin in improving HbA1c. This new stratification might help to tailor and target the imeglimin treatment to patients who would benefit most. Disclosure K.Hagi: Employee; Sumitomo Pharma. K.Kochi: Employee; Sumitomo Dainippon Pharma Co., Ltd., Sumitomo Pharma Co., Ltd. H.Watada: Research Support; Boehringer Ingelheim Japan, Inc., Mitsubishi Tanabe Pharma Corporation, LifeScan Diabetes Institute, Eli Lilly Japan K.K., Sun Pharmaceutical Industries Ltd., Teijin Pharma Limited, Taisho Pharmaceutical Holdings Co., Ltd., Ono Pharmaceutical Co., Ltd., Kowa Company, Ltd., Merck Sharp & Dohme Corp., Sanwa Kagaku Kenkyusho, Speaker's Bureau; Mitsubishi Tanabe Pharma Corporation, Taiho Pharmaceutical Co. Ltd., Novo Nordisk, Abbott Japan Co., Ltd., Astellas Pharma Inc., Merck Sharp & Dohme Corp., Kissei Pharmaceutical Co., Ltd., AstraZeneca, Ono Pharmaceutical Co., Ltd., Boehringer Ingelheim Japan, Inc., Sanofi, Sun Pharmaceutical Industries Ltd., Eli Lilly Japan K.K. K.Kaku: Advisory Panel; Novo Nordisk, Consultant; Sanwa Kagaku Kenkyusho, Speaker's Bureau; Astellas Pharma Inc., Daiichi Sankyo, Eli Lilly Japan K.K., Boehringer Ingelheim Japan, Inc., Kowa Company, Ltd., Mitsubishi Tanabe Pharma Corporation, Taiho Pharmaceutical Co. Ltd., Takeda Pharmaceutical Co., Ltd. K.Ueki: Advisory Panel; Abbott Japan Co., Ltd., Research Support; Abbott Japan Co., Ltd., Boehringer Ingelheim Japan, Inc., Eli Lilly Japan K.K., Sumitomo Dainippon Pharma Co., Ltd., Novo Nordisk, Sanofi K.K., MSD Life Science Foundation, Takeda Pharmaceutical Co., Ltd., Speaker's Bureau; Boehringer Ingelheim Japan, Inc., Eli Lilly Japan K.K., Astellas Pharma Inc., AstraZeneca, Sumitomo Dainippon Pharma Co., Ltd., Novo Nordisk, Taiho Pharmaceutical Co. Ltd., Kowa Company, Ltd., Mitsubishi Tanabe Pharma Corporation, Ono Pharmaceutical Co., Ltd.

Grouping The Regencies/Cities in Indonesia Based on Expenditure Groups Inflation Value Using DBSCAN Method

Inflation is one of the important problems faced by a country in achieving economic goals and targets. The amount of inflation is measured using the Consumer Price Index for eleven expenditure groups in 90 regencies/cities in Indonesia. The occurrence of differences in inflation rates between regencies/cities in Indonesia will affect Indonesia's national inflation. The purpose of this research is to grouping regencies/cities based on expenditure groups inflation value and to identify the characteristics of the resulting groups. DBSCAN is a density-based non-hierarchical cluster method that can be used in data conditions that contain outliers. The data used in this study is secondary data obtained from the publication of the Badan Pusat Statistik Republic of Indonesia (BPS RI) regarding inflation by expenditure group. The analysis includes outlier detection, grouping using the DBSCAN method, performing cluster validation with silhouette coefficient, and identifying the characteristics of the clusters formed. Based on the grouping that has been done, two clusters are produced with a silhouette coefficient value of 0.65. The resulting cluster is cluster 0 in the form of a noise cluster consisting of 3 regencies/cities with regencies/cities that have a high category expenditure group inflation rate. Cluster 1 consisting of 87 regencies/cities is a cluster with regencies/cities that have a low category expenditure group inflation rate.

Human Development Clustering in Indonesia: Using K-Means Method and Based on Human Development Index Categories

The quality of life for Indonesia's population can be measured from the human development index in each province. People who have a good quality of life indicate a prosperous life. The government has the responsibility to advance the welfare of the nation under the mandate of the constitution. The clustering of the human development index (HDI) in Indonesia is used to determine the distribution of quality of life or the distribution of social welfare. In this study, the K-Means method, which is a popular non-hierarchical clustering method, is used to classify human development in each province based on HDI indicators, namely Expected Years of Schooling, Mean Years of Schooling, Adjusted Per Capita Expenditure, and Life Expectancy at Birth. Provinces in Indonesia are clustered into 4 clusters. These results were also compared with the clustering based on HDI categories determined by Statistics Indonesia based on certain cut-off values. According to the HDI category, provinces in Indonesia fall into the medium, high, and very high categories. The results of the two groupings show that there is a trend toward appropriate characteristics for each group. Thus, K-Means can classify provinces in Indonesia according to the characteristics of the HDI indicators.

Pengelompokan Puskesmas Berdasarkan Kasus Balita Stunting di Kabupaten Paser Menggunakan Metode K-Medoids

The number of cases of stunting toddler in Paser Regency increased by 6.66% from 2018 to 2019%. The increased in the number of stunting toddler in Paser Regency shows that the efforts made by the Paser Regency Government have not been effective in reducing the prevalence of stunting toddler because the stunting toddler rate in Paser Regency is still above the threshold set by the World Health Organization (WHO), which is a maximum of 20%. Therefore, an appropriate strategy is needed to find out which areas receive special attention and treatment, one of method to be used is cluster analysis. Cluster analysis is divided into two methods, namely the hierarchical method and the non-hierarchical method. The non-hierarchical method begins by establishing the number of groups. One of the methods included in the non-hierarchical method is K-medoids. In this study, clustering will be carried out in cases of stunting toddlers in Paser Regency using the K-medoids method. This study aims to determine the optimal cluster formed by selecting the smallest Davies Buoldin Index (DBI) value from the 2019 Community Health Center grouping in Paser Regency. The clusters formed for the K-medoids method in this study were 2 clusters, 3 clusters, and 4 clusters. Based on the results of the analysis, the K-medoids method for 2 clusters, 3 clusters and 4 clusters was based on the DBI values ​​of 0.977, 1.470, and 1.670, respectively. The optimal group for classifying stunting toddler cases in Paser Regency in 2019 is 2 cluster using K-medoids method.

NEW STUDENT CLUSTERIZATION BASED ON NEW STUDENT ADMISSION USING DATA MINING METHOD

The process of admitting new students to the Faculty of Information Technology (FTI) at Universitas Budi Luhur produces a large amount of student data in the form of student profile data and other data. This happens causing a buildup of new student data, thus affecting the search for information on that data. This study aims to classify regular undergraduate admissions data at the Faculty of Information Technology (FTI) Universitas Budi Luhur by utilizing the data mining process using the clustering technique. The algorithm used for clustering is the K-Means algorithm. K-Means is a non-hierarchical clustering data method that can group student data into several clusters based on the similarity of the data, so that student data with the same characteristics is grouped in one cluster and those with different characteristics are grouped in another cluster. An implementation using RapidMiner is used to help find accurate values. This research produced a description of what clusters were formed from data on regular undergraduate admissions at the Faculty of Information Technology (FTI) at Universitas Budi Luhur. This will help recommend decision-making to determine the marketing promotion strategy for each study program at Universitas Budi Luhur. Based on the results of the K-Means algorithm cluster, it can also be seen which majors or study programs are of interest in each school from which new students come.

Determination of Seasonal Indices for the Regionalization of Low Flows in the Upper Vistula River Basin

Low flow is a parameter that is used for many purposes in water management. It is particularly important in terms of the needs of small water supply systems, which usually are located in ungauged catchment areas. The aim of this study is to determine the values of seasonality indices for the regionalization of low flow characteristics in the Upper Vistula river basin. This study analyzed three seasonality indicators (seasonality index—SI; seasonality histograms—SHs; and seasonality ratio—SR) in 32 selected catchments differentiated in terms of selected physiographic and meteorological characteristics. Daily flows, from the years 1951–2016, were used in the study. On the basis of the analysis, regions characterized by similar values of seasonality indicators were delineated. The non-hierarchical method of cluster analysis, i.e., the K-means method, was used for this purpose. For each region, obtained by this method, regression models were determined using stepwise regression analysis. Three groups of low flow occurrences were determined: winter, summer and mixed. The delineation of an additional mixed group for the Upper Vistula river basin resulted in a much better fit of calculated flows to observed values than in the case when only two groups of low flow seasonality were determined (winter and summer).

Non-hierarchical work-in-progress control in manufacturing

Reducing Work-In-Process (WIP) in manufacturing systems is associated with advantages such as predictable throughput times and increased manageability. To achieve this, an abundance of WIP control methods have been developed, such as CONWIP11acronym for Constant Work-In-Progress. and Kanban for repetitive manufacturing, and LUMS COR22acronym for Lancaster University Management School Corrected Order Release. and POLCA33acronym for Paired-cell Overlapping Loops of Cards with Authorization. for high-variety manufacturing. These methods take three types of control decisions, viz., release (entry to the manufacturing system), authorization (entry to a work centre) and dispatching (order selection at a work centre). All existing WIP control methods are hierarchical by first deciding on release and authorization before making dispatching decisions, thereby letting the decision whether to produce precede over which order to produce. This hierarchy is traditionally motivated by uncertainty in information between organizational levels, but this is questionable given the advent of Industry 4.0 technologies. We develop a non-hierarchical method – termed DRACO44acronym for Dispatching, Release and Authorization to dynamically Control Order flow. – that simultaneously considers release, authorization and dispatching when deciding. The simulation results show that DRACO significantly outperforms LUMS COR and POLCA on mean WIP and delivery performance measures. Additional analysis also indicates that overall manageability is improved by the non-hierarchical method DRACO.

IMPLEMENTASI ALGORITMA K-MEDOIDS DAN K-ERROR UNTUK PENGELOMPOKAN KABUPATEN/KOTA DI PROVINSI JAWA TENGAH BERDASARKAN JUMLAH PRODUKSI PETERNAKAN TAHUN 2020

The livestock sub-sector is one of the sub-sectors that contribute to the national economy and can significantly absorb labour so that it can be relied upon in efforts to improve the national economy. One of the steps used to increase livestock production in each region in Central Java Province is regional mapping. Cluster analysis is one of the regional mapping methods that can increase livestock production by grouping regencies/cities with characteristics of the same level of livestock production based on the type of livestock production. The k-error and k-medoids method is a non-hierarchical cluster analysis method, where the k-error is a method developed to overcome the problem of data measurement errors in classical cluster analysis, while the k-medoids is a method used to overcome the problem of outliers contained in the data. The validity test of the standard deviation ratio and the WB Index was used to determine the quality of the clustering results. The small validity value of the standard deviation ratio and the WB Index shows the best results of clustering and selecting method. Based on the results of the clustering, the optimal cluster was obtained at k=7 using the k-medoids algorithm, where the validation value of the standard deviation ratio=0.773 and WB Index=0.531.

Climatic regionalization of Montenegro by applying different methods of cluster analysis

To carry out an "objective" regionalization of the climate of Montenegro for the period 1961-2020, this paper used cluster analysis, which is a multivariate technique that classifies a sample of subjects (objects) based on a set of variables into a single number. Based on the results (score), several groups were separated, and similar classes (groups) were grouped into the same cluster. Annual data for mean temperature and total precipitation from 18 meteorological stations were utilized. Temperature and precipitation cluster regions were separated using three different hierarchical agglomerative methods (Unweighted Pair Group Method with Arithmetic Mean (UPGMA), Single linkage, and Ward's) and one non-hierarchical method (K-means). The Euclidean distance was used as a measure of distance for hierarchical methods, and the results were represented graphically in the form of dendrograms and thematic maps. The obtained results indicate that the singled-out temperature and precipitation cluster regions largely coincide with the established climate types in Montenegro. The cluster results further showed that the distribution of meteorological stations clearly reflects the largest part of the climatic diversity of Montenegro and indicates the spatial dimension of temperature and precipitation.

Social infrastructures and socio-economic vulnerability: A socio-territorial integration study in Spanish urban contexts

The growing processes of the socio-economic inequality resulting from the neoliberal political model are evident in large Spanish cities, with a spatial dimension that allows us to visualise its contradictions with great clarity. These social contradictions are particularly marked among the most vulnerable and foreign population groups. For this reason, the study of the spaces and services termed Social Infrastructures (SI), as facilities which favour social integration and cohesion in the city, is of special importance. Our study addresses the distribution of SI in the Spanish cities of Madrid and Barcelona, their relationship with the level of vulnerability in the neighbourhoods and the distribution of the Spanish and foreign population in a novel way, by creating a synthetic SI index and using a non-hierarchical clustering method. Our results show two completely different city models: in Barcelona, SI are distributed following the criterion of socio-territorial equity, while in Madrid, there is a greater presence of SI in the wealthier areas. This leads us to reflect on the social and political implications of our results in both cities.

METHOD OF BUILDING ENSEMBLES OF MODELS FOR DATA CLASSIFICATION BASED ON DECISION CORRELATIONS

The scientific work highlights the problem of increasing the accuracy of binary classification predictions using machine learning algorithms. Over the past few decades, systems that consist of many machine learning algorithms, also called ensemble models, have received increasing attention in the computational intelligence and machine learning community. This attention is well deserved, as ensemble systems have proven to be very effective and extremely versatile in a wide range of problem domains and real-world applications. One algorithm may not make a perfect prediction for a particular data set. Machine learning algorithms have their limitations, so creating a model with high accuracy is a difficult task. If you create and combine several models by combining and aggregating the results of each model, there is a chance to improve the overall accuracy, this problem is dealt with by ensembling. The basis of the information system of binary classification is the ensemble model. This model, in turn, contains a set of unique combinations of basic classifiers – a kind of algorithmic primitives. An ensemble model can be considered as some kind of meta-algorithm, which consists of unique sets of machine learning (ML) classification algorithms. The task of the ensemble model is to find such a combination of basic classification algorithms that would give the highest performance. The performance is evaluated according to the main ML metrics in classification tasks. Another aspect of scientific work is the creation of an aggregation mechanism for combining the results of basic classification algorithms. That is, each unique combination within the ensemble consists of a set of basic models (harbingers), the results of which must be aggregated. In this work, a non-hierarchical clustering method is used to aggregate (average) the predictions of the base models. A feature of this study is to find the correlation coefficients of the base models in each combination. With the help of the magnitude of correlations, the relationship between the prediction of the classifier (base model) and the true value is established, as a result of which space is opened for further research on improving the ensemble model (meta-algorithm)