3164 research outputs found
Sort by
マイクロブログの解析データを利用したWebベースのサファリレビューシス テム開発
Get PDF北見工業大学博士（工学）In this study, I propose the use of online microblogs as review supplements and demonstrate their
applicability through a designed tourist support system that aims to provide additional opinions
and up-to-date points of interest to the less-known tourist spots. In realizing this proposal, I use
Information Extraction (IE), Artificial Intelligence (AI), and Natural Language Processing (NLP)
- based techniques. The proposed approach folds into three.
First, through the use of geotagged tweets. Tweets that contain geolocation information
are considered geotagged and therefore treated as possible tourist on-spot opinions. The main
challenge, however, is to confirm the authenticity of the extracted tweets. This stage includes the
use of location clustering and classification techniques. Specifically, extracted geotagged tweets
are clustered by using location information and then annotated taking into consideration specific
features applied to machine learning-based classification techniques. As for the machine learning
(ML) algorithms, I adopt a fine-tuned transformer neural network-based BERT model which
implements the information of token context orientation for better classification.
Second, I studied geolocatability of ungeotagged tweets so that they can be used as review
alternatives. Ungeotagged tweets have no geolocation information attached so it is difficult to
associate with specific location. Furthermore, Twitter data is typically noisy and consists of
ungrammatical or informal phraseology and non-standard vocabulary, which additionally causes
the feature sparsity problem, resulting in low classifier performance.
To address this, I proposed the use of a two-stage process, a transformer-based model for the
classification of primary tweets, and a combination of impact words like location mention or event
mention for location inferring. Additionally, I evaluate a range of pre-processing techniques for text
categorization to accurately obtain a proper set that collectively contributes to the improvement
of prediction accuracy. A classification framework created here relies on a fine-tuned transformer
neural network model which learns from tweet contents and predicts the locations from which those
tweets were sent - with a limited application in the detection of widely known general locations
- such as tourist spots. I learned that the average 0.84 F1 score of a pre-trained DistilBERT
language model outperformed other tested models when tested on different pre-processing datasets.
Furthermore, i evaluated the effect of impact words like location mention, and event mention on
the geolocation estimation, and model accuracy improvement when impact words are involved or
removed. To investigate the effect of impact words on a classification model, i first computed
the weighting of words using TFIDF and futher created a likelihood wordlist. I discovered model
accuracy improvement as much as 6% when impact words are involved compared to when they
are removed which suggests positive influence of impact words in geolocatability. I also discovered
wrong weighted impact words that negatively contributes to the model performance and byeliminating them, the model F1 score improved by 3%.
Third, I demonstrate the applicability of these two approaches by designing a tourist support
system and mapping extracted opinions to their respective tourist spots as touristic information.doctoral thesi
北見工業大学社会連携推進センター ニュースレター「ウィズ」 No.11
Get PDF■FOCUS
地域循環共生研究推進センター開所式
脱炭素社会の実現に向けて
～カーボンニュートラルセミナー～開催
■TOPIC
「研究室公開2023」開催
社会連携推進センター組織体制
■スタッフ活動othe
Temperature and pressure effects on hydrogen isotope fractionation of methane at a formation of methane hydrate
Get PDFメタンハイドレート生成時，メタンの水素同位体分別によって相対的に軽いメタンがハイドレート相に包接されやすい．本研究では 50 55℃および1.71.7～19.5 MPa の広い温度・圧力範囲でメタンハイドレートを人工的に生成し，メタン水素同位体分別の温度・圧力依存性を詳細に調べた．同位体分別係数は0.98810.9881～0.9932 の範囲にあり，温度依存性は小さいものの，生成圧力に対して増加する，すなわち同位体分別が小さくなる傾向がみられた．このことは，天然ガスハイドレートの包接メタンおよび生成元のメタンが得られれば，これらの水素同位体比の差から結晶生成時の温度・圧力環境が読み取れる可能性を示唆する．journal articl
Random Forest Classifier of Heme Proteins Using PorphyrinDistortions and Axial Ligands of Heme
Get PDFWe achieved success in the construction of a random forest classifier that is capable of accurately classifying heme proteins for oxygen storage/transport function and redox function, with a high level of precision (>83%), by utilizing heme porphyrin distortions and axial ligands of heme as features. Our findings suggest that the heme distortions and the axial ligands are significantly correlated with oxidoreductase functions and oxygen storage/transport functions, respectively, all of which are commonly observed in heme proteins
北海道大津海岸におけるジュエリーアイス出現時期推定手法の一検討
Get PDF冬期に北海道十勝川河口の大津海岸に打ち上げられる「ジュエリーアイス」が新しい観光資源として注目されている．一方で，ジュエリーアイスが見られる時期は，十分には解明されていない．本研究の目的は，ジュエリーアイス出現時期推定手法の開発である．定点カメラで撮影された画像を用いて画像解析を行い，ジュエリーアイスの堆積面積を算出した．出現現象を形成・破壊・輸送・堆積・融解の5つに区分して検討を行った．画像解析によるジュエリーアイスの堆積面積と計算値による堆積量の比較を行い，計算結果の妥当性を確認した．journal articl
Formation mechanism of high-strain bands in commercially pure titanium
Get PDFIn textured α-Ti, band-like high-strain regions at approximately 45° to the loading direction are commonly observed under tensile deformation. Herein, such high-strain regions were called high-strain bands (HSB), and the reasons for the formation mechanism of HSBs were investigated by a uniaxial tensile test and crystal plasticity finite element (CPFE) analyses. First, we conducted a uniaxial tensile test of commercially pure titanium (CP–Ti) with a TD-split texture, where aggregates of (0001) split and incline in the transverse direction, and changes in strain distributions with deformation were observed by digital image correlation. The strain distributions showed that HSBs of approximately 45° to the loading direction were formed from the initial stage of deformation. Second, a geometric model including crystal orientation information was constructed from the crystal orientation map of the CP-Ti specimen obtained by electron back-scattered diffraction, and the cause of the formation of HSBs was investigated by CPFE analysis. The stress–strain relationship and strain distributions obtained by CPFE analysis correlated well with those obtained experimentally, and the HSBs were successfully reproduced. Finally, CPFE analyses were conducted when the loading conditions, critical resolved shear stress (CRSS), or elastic constants were changed. These results show that the HSBs formed from the initial stage of deformation and the distributions formed by elastic deformation changed depending on the elastic anisotropy. However, the distribution of HSBs after plastic deformation was unchanged by elastic anisotropy. This may be because the directions of easy deformation by elastic and plastic deformation coincided. Elastic anisotropy did not affect the distributions of HSBs formed by plastic deformation, and elastic and plastic deformations were predominantly determined by elastic constants and CRSS, respectively. Curved HSBs were also observed when plastic deformation-resistant regions occupied a large area in the specimen. In this case, the distributions of HSBs, caused by elastic deformation did not coincide with those of the plastic deformation.journal articl
A Spatially Integrated Dissolved Inorganic Carbon (SiDIC) Model for Aquatic Ecosystems Considering Submerged Vegetation
Get PDFNet ecosystem production (NEP) by submerged aquatic vegetation plays a substantial role in capturing atmospheric carbon dioxide into aquatic ecosystems. In lakes and estuaries, the net uptake of carbon dioxide by submerged aquatic vegetation is mediated by stratification of the water column which suppresses the vertical flux of carbon dioxide between the upper and lower layers. The presence of submerged aquatic vegetation can also affect the strength of stratification such that the interactions between vegetation, stratification, and NEP can moderate the carbon dioxide emissions. Since stratification can occur in lakes and estuaries, there is need for a new numerical approach able to consider the effect of submerged aquatic vegetation on stratification, NEP, and carbon dioxide. This study aims to develop a model to investigate how stratification, mediated by vegetation density and flexibility, affects the partial pressure of carbon dioxide (pCO2) and dissolved inorganic carbon (DIC). After initial parameterization of coefficients based on experimental work, horizontal and vertical variations in DIC were successfully modeled by a spatially (horizontally) integrated DIC (SiDIC) model, which was validated with field observations from an estuarine and freshwater lake case study. The SiDIC model was able to reproduce the pCO2 changes between daytime and nighttime throughout the water column. Sensitivity tests showed that the fluctuation of pCO2 was controlled by the suppression of stratification due to the density of submerged aquatic vegetation. The results highlight the importance of resolving vegetation-induced stratification when modeling the carbon budget within freshwater lakes and coastal environments.journal articl
Annual Report of Open Facility Center Kitami Institute of Technology Vol.21
Get PDF◆巻頭 共用設備 センター長 大津 直史
◆研究紹介
「核磁気共鳴分光（ NMR ）が必須の糖質化学 」
地球環境工学科 服部 和幸 …1
◆共同利用機器の紹介
「透過電子顕微鏡 (JEM-F200)」
技術部 徳田 奨…7
◆利用実績 …9
◆設置機器類…33
◆活動報告…34
◆編集後記…36othe
Adaptive Optics Microscopy with Wavefront Sensing Based on Neighbor Correlation
Get PDFComplex structures in living cells and tissues induce wavefront errors when light waves pass through them, and images observed with optical microscopes are undesirably blurred. This problem is especially serious for living plant cells because images are strikingly degraded even within a single cell. Adaptive optics (AO) is expected to be a solution to this problem by correcting such wavefront errors, thus enabling high-resolution imaging. In particular, scene-based AO involves wavefront sensing based on the image correlation between subapertures in a Shack–Hartmann wavefront sensor and thus does not require an intense point light source. However, the complex 3D structures of living cells often cause low correlation between subimages, leading to loss of accuracy in wavefront sensing. This paper proposes a novel method for scene-based sensing using only image correlations between adjacent subapertures. The method can minimize changes between subimages to be correlated and thus prevent inaccuracy in phase estimation. Using an artificial test target mimicking the optical properties of a layer of living plant cells, an imaging performance with a Strehl ratio of approximately 0.5 was confirmed. Upon observation of chloroplast autofluorescence inside living leaf cells of the moss Physcomitrium patens, recovered resolution images were successfully obtained even with complex biological structures. Under bright-field illumination, the proposed method outperformed the conventional method, demonstrating the future potential of this method for label- and damage-free AO microscopy. Several points for improvement in terms of the effect of AO correction are discussed.journal articl
氷径分布と空隙率がアイスジャムによる水深上昇に与える影響
Get PDF本研究の目的は，実務での活用を考慮したアイスジャムによる水深上昇の推定手法を開発することである．アイスジャムによる水深上昇の現象を明らかにするために，氷の大きさや厚さ，氷塊の空隙率に着目した二通りの水理実験を実施した．本実験結果より，アイスジャムによる水深上昇の要因は，氷の鉛直下方向への潜り込みと，これに伴う通水面積の減少が推定された．この現象を流水が流れる開口の高さと氷塊の空隙率で表現した実験を実施して実験データを得た．本実験データを基に，流量，川幅，アイスジャム発生前の水深，水の動粘性係数，開口高さ，氷塊の空隙の 6つの入力値から，アイスジャムによる水深上昇の推定手法を開発した．本推定手法による水深比の推定値は，精度高く実験値を推定した．journal articl