Machine Learning-Based Naïve Bayes Classification of Pineapple Productivity: A Case Study in North Sumatra

Abstract. Hasanah Y, Mawarni L, Hanu, H, Lestami A. 2022. Genetic diversity of shallots (Allium ascalonicum L.) from several locations in North Sumatra, Indonesia based on RAPD markers. Biodiversitas 23: 2405-2410. Shallot (Allium ascalonicum L.) is one of the leading horticultural and spice commodities in Indonesia. Assembling shallot varieties to get a good seed quality requires a high genetic diversity which can be analyzed using molecular markers. This study aims to identify the genetic diversity of shallots from several locations in North Sumatra, Indonesia based on Random Amplified Polymorphic DNA (RAPD) markers. This research was conducted at the Biotechnology Laboratory, Faculty of Agriculture, University of Sumatera Utara, Medan from September to October 2021. A total of 11 shallot varieties from several locations in North Sumatra were genetically analyzed based on RAPD markers with OPA-13, OPB-07, OPD-20, OPM-01 primers. The results showed that 11 shallot varieties originating from several locations in North Sumatra had a high genetic diversity with the presence of two main groups at a dissimilarity coefficient of 76%. These beneficial results can be used as a complement to morphological markers in the genetic study of shallots for breeders to decide what genotypes will be crossed to make new genetic combinations in the development of shallots. This study reports the success of the shallot varieties’ fingerprints using RAPD markers.

Strawberry (Fragaria sp.) is primarily grown in temperate and some subtropical countries. With the expansion of fruit commodities in Indonesia, including the introduction of foreign cultivar, strawberry has been increasingly cultivated locally. In North Sumatra, strawberry cultivation, mainly for agritourism, is concentrated in Karo Regency, Berastagi District. This study aimed to detect and identify fungi responsible for wilt disease in strawberry plants across several areas of Berastagi, North Sumatra. This research was conducted from July 2022 to May 2023 at the Plant Disease Laboratory, Faculty of Agriculture, Universitas Sumatera Utara. The study followed Koch’s postulates: the pathogen was isolated and purified from symptomatic plants, then inoculated into healthy plants. Infected plants exhibiting the same symptoms as the initial sample were subsequently re-isolated, purified, and identified at the molecular level. The results confirmed that the causal agent of wilt disease in Daulu (Rini Colia Strawberry, Esy Azera Strawberry) and Dolat Raya (Sonakmalela Strawberry, Alea Strawberry, Sembiring Gurky Strawberry) was Fusarium oxysporum.

Corn (Zea mays L.) is a primary agricultural commodity in Indonesia, alongside rice and soybeans. This study aims to evaluate the growth of local corn plants in North Sumatra after applying various doses of bokashi fertilizer. The research was conducted in Perbaungan Village, Bilah Hulu District, Labuhanbatu Regency, North Sumatra, from December 2024 to February 2025. The study employed a Completely Randomized Design (CRD) with a one-factor experiment, focusing on applying bokashi fertilizer to corn plants. Five treatments were tested: Control (no application), 80 grams per polybag (4 x 20 grams every 2 weeks), 120 grams per polybag (4 x 30 grams every 2 weeks), 180 grams per polybag (4 x 40 grams every 2 weeks), and 200 grams per polybag (4 x 50 grams every 2 weeks). The parameters observed included plant height, number of leaves, and stem diameter. Data analysis was performed using the Tukey Test (BNJ) at a significance level of 5%. The results indicated that the application of bokashi fertilizer significantly influenced the growth of local corn in North Sumatra, with the most favorable outcomes observed at 8 weeks after planting (WAP). The highest plant height recorded was 141.65 cm (P4), the number of leaves was 11.50 strands (P3), and the stem diameter measured 2.73 cm (P4).

Traffic violations have been increasing each year. According to data from the Padang City Police from 2018 to 2023, there were 128,913 traffic violation cases. This is not a small number, and it is time for the police to start utilizing machine learning (ML) technology to evaluate traffic violation cases, as ML can identify hidden patterns or information that cannot be detected manually by conventional statistics or by traffic officers. This research aims to classify traffic violations using the Naïve Bayes algorithm at the Padang City Police by conducting evaluations and comparisons using different dataset ratios. The best algorithm obtained from the comparison will then be analyzed, and the research findings are expected to serve as a reference for the relevant authorities. This research is quantitative in nature, using an experimental method. The data sources or information were obtained from traffic ticket documentation at the Padang City Police and questionnaires distributed to traffic officers of the Padang City Police. The research results show that the Naïve Bayes (NB) algorithm can be used to classify traffic violations at the Padang City Police. The performance test results of the Naïve Bayes (NB) algorithm using all comparison algorithms with different training and testing dataset ratios resulted in 100% accuracy. However, during cross-validation, the Naïve Bayes algorithm achieved the highest accuracy only with training and testing dataset ratios of 80%:20% and 90%:10%. This is due to the large dataset size in this research, which is more than 100,000 entries. The evaluation results of the Naïve Bayes algorithm show that the best model is achieved with the Naïve Bayes algorithm using an 80% training and 20% testing dataset split. Although the performance is similarly high with a 90%:10% training and testing ratio, the researcher chose the 80%:20% training and testing ratio as the best algorithm for reasons of efficiency during training. The argument is that even with just 80%, it is able to predict/classify 20%, which is more efficient than training 90% to predict/classify 10%. Another finding from this implementation is that with a large dataset of 100,000 entries or more, high and stable performance can be achieved, so this research also suggests that to achieve good results from traffic violation classification, the dataset should be above 100,000 entries.

Insolvency prediction and credit rating are challenging tasks used to evaluate the creditworthiness of commercial enterprises based on qualitative and quantitative attributes. One way to approach these tasks is machine learning whereby prediction models are built from sample data. The advantage of machine learning is the automatization of the process obviating the need for human knowledge in most cases and thus, its high level of objectivity. Nevertheless, this approach does not claim to be perfect which is why it does not completely replace human knowledge. Since these models can be used as decision support for experts, interpretable models are desirable. Unfortunately, interpretable models are provided by only a few machine learners. Furthermore, some tasks in finance like credit rating often are multiclass problems. Multiclass classification is often achieved via meta-algorithms using multiple binary learners. However, most state-of-the-art meta-algorithms destroy the interpretability of binary models. In this thesis, we study the performance of interpretable models compared to non-interpretable models in insolvency prediction and credit rating. We look at disjunctive normal forms and decision trees of thresholds of financial ratios as interpretable models. We use random forests, artificial neural networks, and support vector machines as non-interpretable models. Furthermore, we use our own developed machine learning algorithm Thresholder to build disjunctive normal forms and interpretable multiclass models. For the task of insolvency prediction, we demonstrate that interpretable models are not inferior to non-interpretable black-box models. In a first case study, a real-life database with financial statements of 5152 enterprises is used to evaluate the performance for all models. In a second case study focused on credit rating, we show that interpretable multiclass models are even superior to non-interpretable multiclass models. We evaluate their performances on three real-life data sets divided into three rating classes. In these case studies, we compare different interpretable approaches concerning their model size and type of interpretability. We provide example models built on these real-life databases and an interpretation for them. The results show that interpretable threshold-based models are appropriate for classification problems in finance. For these tasks they are not inferior to more sophisticated models like support vector machines. Our algorithm Thresholder builds the smallest models while its performance is comparable to the other interpretable models. In our case study on credit rating, interpretable models perform better than for our case study on insolvency prediction. A possible explanation can be found in the nature of credit rating. In contrast to insolvencies, credit ratings are man-made. This implies that credit ratings are based on decisions by people thinking in interpretable rules, e.g., logical operations on thresholds. Thus, we assume that interpretable models match the problems and detect and represent these interpretable rules.

Cacao is a major agricultural commodity in Indonesia, yet its development is hindered by limited germplasm collections. In this study, the maturase K gene (matK) was used as a marker to determine patterns of genetic variation in Indonesia’s Trinitario and Forastero cacao varieties, with the results showing that the matK sequence does differentiate the varieties. However, the origin of at least one sample is unclear, as it may have been derived from crosses between the Forastero and Trinitario varieties. Similarly, an additional sample appears to be the result of the introduction of Forastero varieties from England, highlighting the importance of careful germplasm collections and molecular studies to identify contaminants.

Coffee is a major agricultural commodity in Indonesia, but its production declined by 1.43% compared to 2021. To boost productivity, vegetative propagation through stem-cutting is a promising method. This study investigates the use of coconut water as a natural plant growth regulator (PGR) to enhance the growth of coffee cuttings. Conducted from June to September 2024 at the Greenhouse of the Faculty of Agriculture, Universitas Pat Petulai, the experiment used a Completely Randomized Design (CRD) with five soaking durations in coconut water: 0 (control), 3, 6, 9, and 12 hours. Results were analyzed using ANOVA and followed by the Least Significant Difference (LSD) test at a 5% significance level. The 9-hour soaking treatment (P3) showed the highest shoot length by the 4th week, while the control had the lowest leaf length. However, soaking duration did not significantly affect shoot diameter, leaf length, or overall shoot growth, indicating a limited influence of coconut water on coffee cutting development

Until now, North Sumatra's food security is still unstable. This condition gives a clear picture that North Sumatra Province experiences problems in the field of food security. Because of the position of the North Sumatra region's very strategic position, this province should be prosperous and independent in terms of food sources. This research aims to develop a policy strategy for the development of food commodities towards national food security. In supporting this research, SWOT, Interpretative Structural Modeling (ISM), and Balanced Score Card (BSC) methods are used. The SWOT method was used to formulate a strategy based on the identification of internal factors and external factors that had a significant influence based on the Mckinsey 7S and PEST models. Furthermore, the ISM method is used to determine strategy priorities. The BSC method is used to determine the implementation plan and mapping of the strategies formulated. The development of food commodity development policy strategies towards food security in North Sumatra based on the results of the research consists of 11 (eleven) main strategies consisting of Strategy SO-1,2,3, Strategy ST-1,2,3, Strategy WO-1,2,3 and Strategy WT-1,2. It is hoped that the eleven strategies resulting from this research can increase the development of food commodities in North Sumatra towards stable and sustainable food security.

Cereals, such as wheat, hold significant nutritional and economic value globally, particularly in regions like Kosovo, where wheat cultivation plays a vital role in the production of staple foods. This study examines three predominant wheat cultivars in Kosovo, namely “Pobeda,” “Euclide,” and “Europe,” focusing on agronomic traits and end product quality to ascertain their suitability for cultivation and subsequent product applications. Through a comprehensive evaluation encompassing agronomic, qualitative, and baking parameters, the study aims to provide insights into cultivar-specific characteristics that influence both cultivation practices and end product quality. Key findings reveal notable disparities among cultivars in agronomic parameters, with “Euclide” demonstrating superior flour extraction rates (42.97%) and exemplary performance in baking evaluations. Noteworthy variations in chemical composition, particularly in fat and fiber content, were observed, with “Pobeda” exhibiting the highest values (2.54% fat content and 3.21% fiber). Despite nonsignificant differences in certain agronomic parameters according to ANOVA results, individual cultivar performance underscored unique traits, highlighting the cultivar-specific influence on overall quality. This research fills a crucial gap by providing a nuanced understanding of wheat cultivars in Kosovo, offering valuable insights for farmers and the baking industry to optimize cultivation techniques and enhance end product quality. The findings advocate for the adoption of cultivar-specific approaches in wheat cultivation and processing, ultimately contributing to sustainable agricultural practices and improved product standards. Overall, this study contributes to the advancement of knowledge in wheat agronomy and quality assessment, facilitating informed decision-making for stakeholders involved in wheat production and processing in Kosovo and beyond.

This study aimed to evaluate the influence of vegetative canopy height on the agronomic characteristics and grape must and wine physicochemical properties of a ‘Cabernet Sauvignon’ vineyard in an espalier-trained system. The evaluated parameters comprised agronomic characteristics of ‘Cabernet Sauvignon’ grapevines and physicochemical compositions of ‘Cabernet Sauvignon’ musts and wines, as well as their phenolic compositions (anthocyanins, stilbenes, and flavonoids), and impact on wine contents of methoxypyrazines (volatile compounds that impart vegetal or earthy odors to wine, which are considered undesirable in large intensity). To that end, four heights of the vegetative canopy were tested: 60 cm (T1), 80 cm (T2), 100 cm (T3), and 120 cm (T4). The experiment was carried out in a commercial vineyard in the region of “Campanha Gaúcha” (Dom Pedrito, Rio Grande do Sul State, Brazil) during the productive cycles of 2015/16, 2016/17, 2017/18, and 2018/19. The main agronomic parameters were measured: estimated productivity per plant and hectare, and mean weight and number of clusters. All wines were elaborated by the same traditional winemaking methods. The physicochemical analyses of must and wines were performed by infrared spectroscopy using Fourier Transform Infrared Spectrometer (FTIR), and the phenolic analysis by high-efficiency liquid chromatography and UV-Vis spectrophotometry. Methoxypyrazines were quantified using headspace solid-phase microextraction (HS-SPME), followed by gas chromatography-mass spectrometry (GC-MS). The results showed that treatments did not influence agronomic parameters. However, technological maturation (sugar accumulation) had interesting results for plants managed at higher canopy heights, with respective results obtained for wine. Treatments had little influence on individual quantification of anthocyanins, although cycles had a high influence on their profile. The wines had low concentrations of methoxypyrazines and did not differ among treatments.

Oil palm, rubber, and cocoa are the top three leading plantation commodities in Indonesia which are usually mutually converted one to each other. Oil palm plantation area and production in Indonesia over the past five years tend to increase, while the cocoa plantation area and production tend to decrease. According to the business ownership status, most of the cocoa plantations are cultivated by smallholders, while most of the oil palm plantations are cultivated by large private plantations. In North Sumatra, the cocoa commodity is the most often converted into oil palm. One of the factors that affect cocoa lands conversion into oil palm is the income prospects. The income analysis and the compare means analysis were used to see whether there are significant differences between the average income of cocoa and oil palm farming. The results showed that there are significant differences between the incomes of cocoa farmers and oil palm farmers, where the prospect of oil palm farmers’ income is higher than the income of cocoa farmers.

Palm oil is the most important plantation commodity in Indonesia. One factor that can suppress oil palm productivity in the field is due to pest attacks, one of which is the attack of the Rhinoceros beetle called Oryctes rhinoceros. This study aims to isolate entomopathogenic fungi found in Oryctes rhinoceros larvae at oil palm plantations in Tanjung Alam Village, Asahan, North Sumatra. The results showed that the fungus isolates had conidia that were greenish and could cause mummification which varied for 7-24 days. The results of molecular identification by PCR (polymerase chain reaction) using ITS1-ITS4 primers showed that the isolates of the entomopathogenic fungi were Metarhizium majus and Metarhizium anisopliae species with DNA bands measuring around 500 base pairs (bp).

Utilization of agricultural products is one of the business opportunities for investors and entrepreneurs. Tea plantations are one of the leading commodities in Indonesia, where one of the producers is North Sumatra, which has three large tea gardens such as Toba Sari, Sidamanik, and Bah Butong. Utilization of tea is only limited to leaves, while tea seeds are not used which unknowingly have the benefit of being able to prevent Golden Snail pests in rice plants. So from that the need for a business that utilizes raw materials that are not worth being of economic value. Further studies and business feasibility studies need to be carried out to determine whether or not the business of organic tea pesticides is feasible in the city of Medan. This research is qualitative and quantitative, by collecting, and analyzing Monik B-Tea business financial data in Medan city. The results of the study can be determined that the results of the analysis show that the feasibility of organic fertilizer business when viewed from a non-financial aspect is feasible. From the technical, the production process uses simple techniques and equipment. From the market aspect, opportunities are still open due to high demand. And from the social aspects of the environment, organic pesticide businesses can contribute to the surrounding community. The analysis was carried out by the cash flow method, Payback Period, Benefit Cost Rational. The results of this analysis showed that the business of organic tea seed pesticides is feasible as shown by the positive final cash balance, the Payback Period which is smaller than the investment age, Cost Rational Benefit is greater than one.

Tangerine or Mandarin orange (Citrus nobilis L) is a well known citrus in the world and becomes a fruit commodity in Indonesia. Various types of local tangerine have been produced with different name depends on the village and Regency where the citrus are grown. However, the information about the genetic variation of the citrus are limited. The research is aimed to study the genetic similarity of the local tangerines by using morphological and simple sequence analysis. The research was carried out by collecting citrus samples from eight villages that are spread in three Regencies at North Sumatera. The morphological characteristic of the citrus are documented, and the DNA are analyze by using simple sequence repeat (SSR) for their genetic variation. The results have revealed that different tangerines (local named as Brastepu, Maga, Sipirok) are having similar morphological characters. The protein in the DNA are containing 48 bands (100 bp-300 bp), consisted of 30 polymorphic bands and 18 monomorphic bands, and have compared to four DNA primers. Analysis of the genetic diversity by using NTsys software found that they are clustered on 0.74 similarity coefficient value and the local citrus are devided into 3 groups. The lowest genetic distance on the Sibanggor Tonga with Baringin Siumuran was 0.63 (63%), meanwhile the highest distance was 1.0 (100%) on Huta Namale with Huta Lombang, Aek Kambiri, and Aek Horsik..

Rubber is a key plantation commodity in Indonesia, serving as a significant source of non-oil and gas foreign exchange. However, rubber productivity remains low, largely due to inadequate cultivation technology and the impact of pests and diseases. One of the most economically significant diseases affecting rubber plants is deciduous disease caused by Colletotrichum gloeosporioides, which spreads through spores carried by wind and rain. This study was conducted at the Sungei Putih Rubber Research Institute, North Sumatra, from February to June 2019, at an elevation of ± 25 meters above sea level. A factorial randomized block design (RBD) with three replications was employed. The treatments included three types of endophytic fungi (E1, E2, E3) derived from different isolates of PB 260 rubber clones, and four application methods (M0, M1, M2, M3). The observed parameters included latent period, disease occurrence, and disease intensity. The results indicated that the E2M2 treatment, which combined the second endophytic fungus with metabolite application, was the most effective in controlling C. gloeosporioides deciduous disease. This combination resulted in the highest average latent period (3.67%), the lowest disease occurrence (0.03%), and the most effective reduction in disease intensity (17.67% after 12 days). The study demonstrates the potential of specific endophytic fungi and application methods in managing rubber plant diseases, offering insights for improving rubber productivity in Indonesia.