| Information analysis processing is the observer in any detectable way wheneverthe information changes (processing). Therefore, it can describe all the changes in theworld, arrange from falling from rock to print a text in a digital computing system.American mathematician Claude Elwood Shannon, who gives the information theory,had given the information analysis process a clear definition: Conversion to clearinformation from the hidden information. The information analysis process, as thebasic needs of human survival and development, penetrate into the human activities inall directions. In this paper, the study is forcing on two aspects, gene expression dataanalysis and touch-based device information analysis, which belongs to predictiontype and comparison type information analysis respectively.Case one: the study on gene expression data analysis. The study on the diseaseresistance related gene is important to the production of rice. Due to the similar genewould have similar gene expression, using the DNA microarray data to study thefunction of the gene is meaningful. In this paper, an improved Support VectorMachine Recursive Feature Extraction (SVM-RFE) algorithm was used to screen thedisease resistance genes. In this study the gene expression data is from the NCBI GEOpublic database, and choose the GSE16142as disease resistance data research sample,and this series contains two sub-series GSE16140and GSE16141, GSE16140namedTW16, belongs to Rice tungro disease of resistant varieties; GSE16141contains12samples named Oryzasativa cv. Taichung the Native1, belongs to rice tungro diseasesusceptible varieties. In the study, firstly, we do the T-test; secondly, normalized thedata to [-1,1] interval; thirdly, using the linear function to train the data set. In order tosolve the problem of instability of SVM-RFE algorithm, the algorithm process wasimproved, leave-one-out cross validation (LOOCV) was used to eliminate the features21times. In the20important genes we found that3of them have strong relation tothe disease resistance as reported and2of them have relation to the stress response. Itshows that our method can find out which genes could impact the rice growth status(normal/disease) and might provide a guide on finding other unknown rice disease resistance/sensibility genes in biology.Case two: the study on touch-based device information analysis. The touch-baseddevice is more and more popular recently, most of the major manufacturer of mobiledevices in the world have touch-based product. Thus, this study is to evaluate theperformance of ring scrolling and flicking scrolling techniques for documentnavigation in touch-based mobile devices. In this paper, we use Latin square tobalance the result of the experiment. The experiment was a within-subjects designwith three input methods: thumb, index finger, and pen input. The two techniques tobe compared in the experiment (flicking scrolling and ring scrolling) were counterbalanced. Each block consisted of6trials while each trial contained9combinations(distance of target line (3) x frame width (3)). Each combination consisted of tenrepeated phases. There are twelve participants (11males,1female and average age25). After remove the learning effect from the raw data, we analyses the result byANOVA. Experimental results showed that flicking scrolling performed better thanring scrolling for each of three input methods regardless of the distance between twotargets or the width of the frame. The study also implied that simple MT model givenby Anderson in2005is more suitable for human performance on touch-based device.This study provides an insight for document navigation UI design in touch-basedmobile devices. |
PDF Full Text Request