• Title/Summary/Keyword: 가중치분석

Search Result 2,161, Processing Time 0.031 seconds

Development of an Eating Habit Checklist for Screening Elementary School Children at Risk of Inadequate Micronutrient Intake (초등학생의 미량영양소 섭취부족 위험 진단을 위한 간이 식습관평가표 개발)

  • Yon, Mi-Yong;Hyun, Tai-Sun
    • Journal of Nutrition and Health
    • /
    • v.42 no.1
    • /
    • pp.38-47
    • /
    • 2009
  • The purpose of the study was to develop an eating habit checklist for screening elementary school children at risk of inadequate micronutrient intake. Eating habits, food intake, and anthropometric data were collected from 142 children (80 boys and 62 girls) in the $4^{th}$ to $6^{th}$ grades of elementary schools. Percentage of Recommended Intakes (RI) and Mean Adequacy Ratio (MAR) of six micronutrients; vitamin A, riboflavin, vitamin C, calcium, iron, zinc, and the number of nutrients the children consumed below EAR among the six nutrients were used as indices to detect the risk of inadequate micronutrient intake. Pearson correlation coefficients were calculated between eating habit scores and inadequate micronutrient intake indices in order to select questions included in the checklist. Meal frequency, enough time for breakfast, regularity of dinner, appetite, eating frequencies of Kimchi, milk, fruits and beans showed significant correlations with indices of inadequate micronutrient intake. Stepwise regression analysis was performed to give each item a different weight by prediction strength. To determine the cut-off point of the test score, sensitivity, specificity, and positive predictive values were calculated. The 8-item checklist with test results from 0 to 12 points was developed, and those with equal or higher than 6 points were diagnosed as high-risk group of inadequate micronutrient intake, and those with 4 or 5 points were diagnosed as moderate-risk group. Among our subjects 14.1% was diagnosed as high-risk group, and 30.3% as moderate-risk group. The proportions of the subjects who consumed below EAR of all micronutrients but vitamin C were highest in the high-risk group, and there were significant differences in the proportions of the subjects with intake below EAR of all micronutrients except vitamin B6 among the three groups. This checklist will provide a useful screening tool to identify children at risk of inadequate micronutrient intake.

Solid Waste Disposal Site Selection in Rural Area: Youngyang-Gun, Kyungpook (농촌지역 쓰레기 매립장 입지선정에 관한 연구 -경상북도 영양군을 사례로-)

  • Park, Soon-Ho
    • Journal of the Korean association of regional geographers
    • /
    • v.3 no.1
    • /
    • pp.63-80
    • /
    • 1997
  • This study attempts to establish the criteria of site selection for establishing solid waste disposal facility, to determine optimal solid waste disposal sites with the criteria, and to examine the suitability of the selected sites. The Multi-Criteria Evaluation(MCE) module in Idrisi is used to determine optimal sites for solid waste disposal. The MCE combines the information from several criteria in interval and/or ratio scale to form a single index of evaluation without leveling down the data scale into ordinal scale. The summary of this study is as follows: First, the considerable criteria are selected through reviewing the literature and the availability of data: namely, percent of slope, fault lines, bedrock characteristics, major residential areas, reservoirs of water supply, rivers, inundated area, roads, and tourist resorts. Second, the criteria maps of nine factors have been developed. Each factor map is standardized and multiplies by its weight, and then the results are summed. After all of the factors have been incorporated, the resulting suitability map is multiplied by each of the constraint in turn to "zero out" unsuitable area. The unsuitable areas are discovered in urban district and its adjacencies, and mountain region as well as river, roads, resort area and their adjacency districts. Third, the potential sites for establishing waste disposal facilities are twenty five districts in Youngyang-gun. Five districts are located in Subi-myun Sinam-ri, nine districts in Chunggi-myun Haehwa-ri and Moojin-ri, and eleven districts in Sukbo-myun Posan-ri. The first highest score of suitability for waste disposal sites is shown at number eleven district in Chunggi-myun Moojin-ri and the second highest one is discovered at number twenty one district in Sukbo-myun Posan-ri that is followed by number nine district in Chunggi-myun Haehwa-ri, number seventeen and twenty three in Sukbo-myun Posan-ri, and number two in Subi-myun Sinam-ri. The first lowest score is found in number six district in Chunggi-myun Haehwa-ri, and the second lowest one is number five district in Subi-myun Sinam-ri. Finally, the Geographic Information System (GIS) helps to select optimal sites with more objectively and to minimize conflict in the determination of waste disposal sites. It is important to present several potential sites with objective criteria for establishing waste disposal facilities and to discover characteristics of each potential site as a result of that final sites of waste disposal are determined through considering thought of residents. This study has a limitation of criteria as a result of the restriction of availability of data such as underground water, soil texture and mineralogy, and thought of residents. To improve selection of optimal sites for a waste disposal facility, more wide rage of spatial and non-spatial data base should be constructed.

  • PDF

Detection of Pine Wilt Disease tree Using High Resolution Aerial Photographs - A Case Study of Kangwon National University Research Forest - (시계열 고해상도 항공영상을 이용한 소나무재선충병 감염목 탐지 - 강원대학교 학술림 일원을 대상으로 -)

  • PARK, Jeong-Mook;CHOI, In-Gyu;LEE, Jung-Soo
    • Journal of the Korean Association of Geographic Information Studies
    • /
    • v.22 no.2
    • /
    • pp.36-49
    • /
    • 2019
  • The objectives of this study were to extract "Field Survey Based Infection Tree of Pine Wilt Disease(FSB_ITPWD)" and "Object Classification Based Infection Tree of Pine Wilt Disease(OCB_ITPWD)" from the Research Forest at Kangwon National University, and evaluate the spatial distribution characteristics and occurrence intensity of wood infested by pine wood nematode. It was found that the OCB optimum weights (OCB) were 11 for Scale, 0.1 for Shape, 0.9 for Color, 0.9 for Compactness, and 0.1 for Smoothness. The overall classification accuracy was approximately 94%, and the Kappa coefficient was 0.85, which was very high. OCB_ITPWD area is approximately 2.4ha, which is approximately 0.05% of the total area. When the stand structure, distribution characteristics, and topographic and geographic factors of OCB_ITPWD and those of FSB_ITPWD were compared, age class IV was the most abundant age class in FSB_ITPWD (approximately 55%) and OCB_ITPWD (approximately 44%) - the latter was 11% lower than the former. The diameter at breast heigh (DBH at 1.2m from the ground) results showed that (below 14cm) and (below 28cm) DBH trees were the majority (approximately 93%) in OCB_ITPWD, while medium and (more then 30cm) DBH trees were the majority (approximately 87%) in FSB_ITPWD, indicating different DBH distribution. On the other hand, the elevation distribution rate of OCB_ITPWD was mostly between 401 and 500m (approximately 30%), while that of FSB_ITPWD was mostly between 301 and 400m (approximately 45%). Additionally, the accessibility from the forest road was the highest at "100m or less" for both OCB_ITPWD (24%) and FSB_ITPWD (31%), indicating that more trees were infected when a stand was closer to a forest road with higher accessibility. OCB_ITPWD hotspots were 31 and 32 compartments, and it was highly distributed in areas with a higher age class and a higher DBH class.

A Study on the Measurement of Startup and Venture Ecosystem Index (창업·벤처 생태계 측정에 관한 연구)

  • Kim, Sunwoo;Jin, Wooseok;Kwak, Kihyun;Ko, Hyuk-Jin
    • Asia-Pacific Journal of Business Venturing and Entrepreneurship
    • /
    • v.16 no.6
    • /
    • pp.31-42
    • /
    • 2021
  • The importance of startups and ventures in the Korean economy is growing. This study measured whether the start-up and venture ecosystem is growing, including the growth of startups and ventures. The startup and venture ecosystem consists of startups and ventures, investors, and government, which are the main actors of the 'ecosystem', and their movements were measured with 25 quantitative indicators. Based on the original data of the time series from 2010 to 2020, the startup and venture ecosystem index was calculated by applying weights through the comprehensive stock index method and AHP. In 2020, the startup and venture ecosystem grew 2.9 times compared to 2010, and the increase in the government index had a significant impact on growth. Also, the individual indicators that make up each index in 2020, the corporate index had the greatest impact on the growth of the number of 100-billion ventures, while the investment index had a recovery amount and the government index had a significant impact. Based on the original data, the startup and venture ecosystem index was analyzed by dividing it into ecosystems (startup ecosystem and venture ecosystem), industry by industry (all industries and manufacturing industry), and region (Korea and Busan). As a result, the growth of the startup ecosystem over the past decade has been slightly larger than that of the venture ecosystem. The manufacturing was lower than that of all industries, and Busan was lower than that of the nation. This study was intended to use it for the establishment and implementation of support policies by developing, measuring, and monitoring the startup and venture ecosystem index. This index has the advantage of being able to research the interrelationships between major actors, and anyone can calculate the index using the results of official statistical surveys. In the future, it is necessary to continuously update this content to understand how economic and social events or policy support have affected the startup and venture ecosystem.

Comparison of ESG Evaluation Methods: Focusing on the K-ESG Guideline (ESG 평가방법 비교: K-ESG 가이드라인을 중심으로)

  • Chanhi Cho;Hyoung-Yong Lee
    • Journal of Intelligence and Information Systems
    • /
    • v.29 no.1
    • /
    • pp.1-25
    • /
    • 2023
  • ESG management is becoming a necessity of the times, but there are about 600 ESG evaluation indicators worldwide, causing confusion in the market as different ESG ratings were assigned to individual companies according to evaluation agencies. In addition, since the method of applying ESG was not disclosed, there were not many ways for companies that wanted to introduce ESG management to get help. Accordingly, the Ministry of Trade, Industry and Energy announced the K-ESG guideline jointly with the ministries. In previous studies, there were few studies on the comparison of evaluation grades by ESG evaluation company or the application of evaluation diagnostic items. Therefore, in this study, the ease of application and improvement of the K-ESG guideline was attempted by applying the K-ESG guideline to companies that already have ESG ratings. The position of the K-ESG guideline is also confirmed by comparing the scores calculated through the K-ESG guideline for companies that have ESG ratings from global ESG evaluation agencies and domestic ESG evaluation agencies. As a result of the analysis, first, the K-ESG guideline provide clear and detailed standards for individual companies to set their own ESG goals and set the direction of ESG practice. Second, the K-ESG guideline is suitable for domestic and global ESG evaluation standards as it has 61 diagnostic items and 12 additional diagnostic items covering the evaluation indicators of global representative ESG evaluation agencies and KCGS in Korea. Third, the ESG rating of the K-ESG guideline was higher than that of a global ESG rating company and lower than or similar to that of a domestic ESG rating company. Fourth, the ease of application of the K-ESG guideline is judged to be high. Fifth, the point to be improved in the K-ESG guideline is that the government needs to compile industry average statistics on diagnostic items in the K-ESG environment area and publish them on the government's ESG-only site. In addition, the applied weights of E, S, and G by industry should be determined and disclosed. This study will help ESG evaluation agencies, corporate management, and ESG managers interested in ESG management in establishing ESG management strategies and contributing to providing improvements to be referenced when revising the K-ESG guideline in the future.

Metabolic risk and nutritional state according to breakfast energy level of Korean adults: Using the 2007~2009 Korea National Health and Nutrition Examination Survey (한국 성인의 아침식사 에너지 수준에 따른 대사적 위험과 영양상태: 2007~2009년 국민건강영양조사 자료 이용)

  • Jang, So-Hyoun;Suh, Yoon Suk;Chung, Young-Jin
    • Journal of Nutrition and Health
    • /
    • v.48 no.1
    • /
    • pp.46-57
    • /
    • 2015
  • Purpose: The aim of this study was to determine an appropriate energy level of breakfast with less risk of chronic disease for Korean adults. Methods: Using data from the 2007~2009 Korean National Health & Nutrition Examination Survey, from a total of 12,238 adults aged 19~64, the final 7,769 subjects were analyzed except subjects who were undergoing treatment for cancer or metabolic disorder. According to the percent of breakfast energy intake versus their estimated energy requirement (EER), the subjects were divided into four groups: < 10% (very low, VL), 10~20% (low, L), 20~30% (moderate, M), ${\geq}30%$ (sufficient, S). All data were analyzed on the metabolic risk and nutritional state after application of weighted value and adjustment of sex, age, residential area, income, education, job or jobless, and energy intake using a general linear model or logistic regression. Results: The subjects of group S were 16.9% of total subjects, group M 39.2%, group L 37.6%, and group VL 6.3%. The VL group included more male subjects, younger-aged (19 to 40 years), urban residents, higher income, higher education, and fewer breakfasts eaters together with family members. Among the 4 groups, the VL group showed the highest waist circumference, while the S group showed the lowest waist circumference, body mass index, and serum total cholesterol. The groups of VL and L with lower intake of breakfast energy showed high percent of energy from protein and fat, and low percent of energy from carbohydrate. With the increase of breakfast energy level, intake of energy, most nutrients and food groups increased, and the percentage of subjects consuming nutrients below EAR decreased. The VL group showed relatively higher intake of snacks, sugar, meat and eggs, oil, and seasonings, and the lowest intake of vegetable. Risk of obesity by waist circumference was highest in the VL group by 1.90 times of the S group and the same trend was shown in obesity by BMI. Risk of dyslipidemia by serum total cholesterol was 1.84 times higher in the VL group compared to the S group. Risk of diabetes by Glu-FBS (fasting blood sugar) was 1.57 times higher in the VL group compared to the S group. Conclusion: The results indicate that higher breakfast energy level is positively related to lower metabolic risk and more desirable nutritional state in Korean adults. Therefore, breakfast energy intake more than 30% of their own EER would be highly recommended for Korean adults.

Development of a split beam transducer for measuring fish size distribution (어체 크기의 자동 식별을 위한 split beam 음향 변환기의 재발)

  • 이대재;신형일
    • Journal of the Korean Society of Fisheries and Ocean Technology
    • /
    • v.37 no.3
    • /
    • pp.196-213
    • /
    • 2001
  • A split beam ultrasonic transducer operating at a frequency of 70 kHz to use in the fish sizing echo sounder was developed and the acoustic radiation characteristics were experimentally analyzed. The amplitude shading method utilizing the properties of the Chebyshev polynomials was used to obtain side lobe levels below -20 dB and to optimize the relationship between main beam width and side lobe level of the transducer, and the amplitude shading coefficient to each of the elements was achieved by changing the amplitude contribution of elements with 4 weighting transformers embodied in the planar array transducer assembly. The planar array split beam transducer assembly was composed of 36 piezoelectric ceramics (NEPEC N-21, Tokin) of rod type of 10 mm in diameter and 18.7 mm in length of 70 kHz arranged in the rectangular configuration, and the 4 electrical inputs were supplied to the beamformer. A series of impedance measurements were conducted to check the uniformity of the individual quadrants, and also in the configurations of reception and transmission, resonant frequency, and the transmitting and receiving characteristics were measured in the water tank and analyzed, respectively. The results obtained are summarized as follows : 1. Average resonant and antiresonant frequencies of electrical impedance for four quadrants of the split beam transducer in water were 69.8 kHz and 83.0 kHz, respectively. Average electrical impedance for each individual transducer quadrant was 49.2$\Omega$ at resonant frequency and 704.7$\Omega$ at antiresonant frequency. 2. The resonance peak in the transmitting voltage response (TVR) for four quadrants of the split beam transducer was observed all at 70.0 kHz and the value of TVR was all about 165.5 dB re 1 $\mu$Pa/V at 1 m at 70.0 kHz with bandwidth of 10.0 kHz between -3 dB down points. The resonance peak in the receiving sensitivity (SRT) for four combined quadrants (quad LU+LL, quad RU+RL, quad LU+RU, quad LL+RL) of the split beam transducer was observed all at 75.0 kHz and the value of SRT was all about -177.7 dB re 1 V/$\mu$Pa at 75.0 kHz with bandwidth of 10.0 kHz between -3 dB down points. The sum beam transmitting voltage response and receiving senstivity was 175.0 dB re 1$\mu$Pa/V at 1 m at 75.0 kHz with bandwidth of 10.0 kHz, respectively. 3. The sum beam of split beam transducer was approximately circular with a half beam angle of $9.0^\circ$ at -3 dB points all in both axis of the horizontal plane and the vertical plane. The first measured side lobe levels for the sum beam of split beam transducer were -19.7 dB at $22^\circ$ and -19.4 dB at $-26^\circ$ in the horizontal plane, respectively and -20.1 dB at $22^\circ$ and -22.0 dB at $-26^\circ$ in the vertical plane, respectively. 4. The developed split beam transducer was tested to estimate the angular position of the target in the beam through split beam phase measurements, and the beam pattern loss for target strength corrections was measured and analyzed.

  • PDF

Bone mineral density and nutritional state according to milk consumption in Korean postmenopausal women who drink coffee: Using the 2008~2009 Korea National Health and Nutrition Examination Survey (한국 폐경 후 여성 커피소비자에서 우유섭취여부에 따른 골밀도와 영양상태 비교 : 2008~2009년 국민건강영양조사 자료 이용)

  • Ryu, Sun-Hyoung;Suh, Yoon Suk
    • Journal of Nutrition and Health
    • /
    • v.49 no.5
    • /
    • pp.347-357
    • /
    • 2016
  • Purpose: This study investigated bone mineral density and nutritional state according to consumption of milk in Korean postmenopausal women who drink coffee. Methods: Using the 2008~2009 Korean National Health & Nutrition Examination Survey data, a total of 1,373 postmenopausal females aged 50 yrs and over were analyzed after excluding those with diseases related to bone health. According to coffee and/or milk consumption, subjects were divided into four groups: coffee only, both coffee & milk, milk only, and none of the above. All data were processed after application of weighted values and adjustment of age, body mass index, physical activity, drinking, and smoking using a general linear model. For analysis of nutrient intake and bone density, data were additionally adjusted by total energy and calcium intake. Results: The coffee & milk group had more subjects younger than 65 yrs and higher education, urban residents, and higher income than any other group. The coffee only group showed somewhat similar characteristics as the none of the above group, which showed the highest percentage of subjects older than 65 and in a lower education and socio-economic state. Body weight, height, body mass index, and lean mass were the highest in coffee & milk group and lowest in the none of the above group. On the other hand, the milk only group showed the lowest values for body mass index and waist circumference, whereas percent body fat did not show any difference among the groups. The coffee and milk group showed the highest bone mineral density in the total femur and lumbar spine as well as the highest nutritional state and most food group intakes, followed by the milk only group, coffee only group, and none of the above group. In the assessment of osteoporosis based on T-score of bone mineral density, although not significant, the coffee and milk group and milk only group, which showed a better nutritional state, included more subjects with a normal bone density, whereas the none of the above group included more subjects with osteoporosis than any other group. Conclusion: Bone mineral density in postmenopausal women might not be affected by coffee drinking if their diets are accompanied by balanced food and nutrient intake including milk.

A Study on Searching for Export Candidate Countries of the Korean Food and Beverage Industry Using Node2vec Graph Embedding and Light GBM Link Prediction (Node2vec 그래프 임베딩과 Light GBM 링크 예측을 활용한 식음료 산업의 수출 후보국가 탐색 연구)

  • Lee, Jae-Seong;Jun, Seung-Pyo;Seo, Jinny
    • Journal of Intelligence and Information Systems
    • /
    • v.27 no.4
    • /
    • pp.73-95
    • /
    • 2021
  • This study uses Node2vec graph embedding method and Light GBM link prediction to explore undeveloped export candidate countries in Korea's food and beverage industry. Node2vec is the method that improves the limit of the structural equivalence representation of the network, which is known to be relatively weak compared to the existing link prediction method based on the number of common neighbors of the network. Therefore, the method is known to show excellent performance in both community detection and structural equivalence of the network. The vector value obtained by embedding the network in this way operates under the condition of a constant length from an arbitrarily designated starting point node. Therefore, it has the advantage that it is easy to apply the sequence of nodes as an input value to the model for downstream tasks such as Logistic Regression, Support Vector Machine, and Random Forest. Based on these features of the Node2vec graph embedding method, this study applied the above method to the international trade information of the Korean food and beverage industry. Through this, we intend to contribute to creating the effect of extensive margin diversification in Korea in the global value chain relationship of the industry. The optimal predictive model derived from the results of this study recorded a precision of 0.95 and a recall of 0.79, and an F1 score of 0.86, showing excellent performance. This performance was shown to be superior to that of the binary classifier based on Logistic Regression set as the baseline model. In the baseline model, a precision of 0.95 and a recall of 0.73 were recorded, and an F1 score of 0.83 was recorded. In addition, the light GBM-based optimal prediction model derived from this study showed superior performance than the link prediction model of previous studies, which is set as a benchmarking model in this study. The predictive model of the previous study recorded only a recall rate of 0.75, but the proposed model of this study showed better performance which recall rate is 0.79. The difference in the performance of the prediction results between benchmarking model and this study model is due to the model learning strategy. In this study, groups were classified by the trade value scale, and prediction models were trained differently for these groups. Specific methods are (1) a method of randomly masking and learning a model for all trades without setting specific conditions for trade value, (2) arbitrarily masking a part of the trades with an average trade value or higher and using the model method, and (3) a method of arbitrarily masking some of the trades with the top 25% or higher trade value and learning the model. As a result of the experiment, it was confirmed that the performance of the model trained by randomly masking some of the trades with the above-average trade value in this method was the best and appeared stably. It was found that most of the results of potential export candidates for Korea derived through the above model appeared appropriate through additional investigation. Combining the above, this study could suggest the practical utility of the link prediction method applying Node2vec and Light GBM. In addition, useful implications could be derived for weight update strategies that can perform better link prediction while training the model. On the other hand, this study also has policy utility because it is applied to trade transactions that have not been performed much in the research related to link prediction based on graph embedding. The results of this study support a rapid response to changes in the global value chain such as the recent US-China trade conflict or Japan's export regulations, and I think that it has sufficient usefulness as a tool for policy decision-making.

Transfer Learning using Multiple ConvNet Layers Activation Features with Principal Component Analysis for Image Classification (전이학습 기반 다중 컨볼류션 신경망 레이어의 활성화 특징과 주성분 분석을 이용한 이미지 분류 방법)

  • Byambajav, Batkhuu;Alikhanov, Jumabek;Fang, Yang;Ko, Seunghyun;Jo, Geun Sik
    • Journal of Intelligence and Information Systems
    • /
    • v.24 no.1
    • /
    • pp.205-225
    • /
    • 2018
  • Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.