Difference between revisions of "Team:NCTU Formosa/Model"

 
Line 303: Line 303:
  
 
     .seemore{
 
     .seemore{
       width: 12%;
+
       width: 15%;
       margin-left: 44%;
+
       margin-left: 42.5%;
 
       cursor: pointer;
 
       cursor: pointer;
 
       transition-duration: 0.4s;
 
       transition-duration: 0.4s;
Line 314: Line 314:
  
 
     .close{
 
     .close{
       width: 12%;
+
       width: 15%;
       margin-left: 44%;
+
       margin-left: 42.5%;
 
       cursor: pointer;
 
       cursor: pointer;
 
       transition-duration: 0.4S;
 
       transition-duration: 0.4S;
Line 336: Line 336:
 
       padding: 1vw;
 
       padding: 1vw;
 
       padding-bottom: 3vw;
 
       padding-bottom: 3vw;
 +
      padding-top: 3vw;
 
       width: 80%;
 
       width: 80%;
 
       margin-left: 10%;
 
       margin-left: 10%;
Line 453: Line 454:
 
     </div>
 
     </div>
 
     <div class="assumption">
 
     <div class="assumption">
      <img src="https://static.igem.org/mediawiki/2018/0/0d/T--NCTU_Formosa--microbiota.png" class="assumption_icon">
 
 
       <div class="title444"><p>Assumptions of Whole Model</p></div>
 
       <div class="title444"><p>Assumptions of Whole Model</p></div>
       <div class="text" style="width: 85%; margin-left: 7.5%;">
+
       <div class="text" style="width: 85%; margin-left: 7.5%; margin-bottom: 2vw;">
 
         <p>
 
         <p>
 
           &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;The interactions in the nature is very complicated. If we take all the conditions into account in our models, they will be too complicated to compute. Thus, according to our human practices and researches, we set several assumptions to simplify our model. However, our models still achieved high accuracy with the power of machine learning. Please click the button below to see the assumptions!
 
           &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;The interactions in the nature is very complicated. If we take all the conditions into account in our models, they will be too complicated to compute. Thus, according to our human practices and researches, we set several assumptions to simplify our model. However, our models still achieved high accuracy with the power of machine learning. Please click the button below to see the assumptions!

Latest revision as of 16:53, 7 December 2018

Navigation Bar Template

Overview

     The purpose of our smart farming system is to precisely regulate soil microbiota using bio-stimulators to achieve a desired effect. While microbiology and ecology drive the efficiency of our bio-stimulators, dry-lab analyses and models power the control center of our system. To truly achieve precise regulation, we designed a system of interconnected models, linked sensors together through our IoTtalk platform and strengthened by continuous feedback of data. NGS data provides invaluable details of our bacterial regulation network. Machine learning software in the form of Weka processes these details, granting prediction capabilities made more accurate through self-learning. An electrical conductivity sensor details levels of nitrogen, phosphorus and potassium present in soil and alerts farmers when another application is needed, and a curcumin sensor allows for consistent monitoring of curcumin concentrations without damaging plants. Both sensors transmit figures constantly through IoTtalk, providing steady data for calibration of their respective models through artificial intelligence. Finally, inhibition modelling of newly constructed bio-stimulators in the form of bacteriocins grants even greater precision, while peptide prediction using the Scoring Card Method characterizes even more novel and efficient bio-stimulators. Click below to learn how we turn farming into a science!

Microbiota Prediction


Predict how biostimulators effect bacterial distribution using Weka machine learning.

See More>>

Peptide Prediction


Scoring Card Method predicts new antimicrobial peptides as new bio-stimulators.

See More>>

Growth Model


Relate soil factors and bacteriocins affect B. subtilis growth through Simulink.

See More>>

Productivity Model


Model the relationship between cumulative fertilizer use and final productivity.

See More>>

NGS Data Analysis


Analyze the microbiota and ensure soil health with next generation sequencing.

See More>>

Assumptions of Whole Model

     The interactions in the nature is very complicated. If we take all the conditions into account in our models, they will be too complicated to compute. Thus, according to our human practices and researches, we set several assumptions to simplify our model. However, our models still achieved high accuracy with the power of machine learning. Please click the button below to see the assumptions!

Microbiota Prediction

1. The growth of the bacteria and the correlation between two different bacteria are only relative with N, P, and K fertilizer in soil.

2. Soil organic carbon is constant.

3. Unknown bacteria are not in the soil.

4. N, P, K will not loss after spreading in the soil.

5. The growth rate of all bacteria has same relationship with soil organic carbon.

6. The function of bacteria will not change with the environment.

7. If the spearman's rank correlation coefficient is lower than -0.7 or higher than 0.7, our model will take this correlationship into account.

Growth Model

1. Temperature, pH, EC is constant after we detect first time.

2. The growth of B. sub only depends on the conditions of temperature, pH, EC, N, P, and K.

3. The growth of B. sub doesn’t interfere with other bacteria.

4. The degradation of bacteriocin doesn’t interfere with external factors.

5. After the microbiota reaches the dynamic equilibrium, it will only affected by bio-stimulator.

Productivity Model

1. The growth of plant is only influence on N, P, and K in soil.

2. After spreading fertilizer, plant will absorb all of it before we spread fertilizer next time.

NGS Data Analysis

1. The lower bacteria ratio is, the smaller influence it causes.