Agriculture and Livestock Informatics Research. Malik Jahan Khan, PhD Associate Professor (Computer Science) Namal College, Mianwali
|
|
- Jerome Caldwell
- 5 years ago
- Views:
Transcription
1 Agriculture and Livestock Informatics Research Malik Jahan Khan, PhD Associate Professor (Computer Science) Namal College, Mianwali
2 Namal s Vision To become center of academic excellence for rural uplift through: Educating bright youth Solving rural problems through innovation 2
3 Some Facts ~70% of Pakistan is rural ~24% of Gross Domestic Product (GDP) comes from agriculture sector ~49% of total workforce in Pakistan is employed by agriculture sector ~65% of population is directly or indirectly linked with agriculture ~27% of total area of the country is under cultivation Source: Economic survey: Govt. of Pakistan 3
4 Some Facts Pakistan is: 4 th largest cotton producing country in the world 7 th largest wheat producing country in the world Agriculture GDP has declined over the last decade Total agriculture production is almost 50% of its potential Sources: FAO of United Nations, Ministry of Finance (Govt. of Pakistan) 4
5 Intelligence Disease Diagnosis of Crops Lack of availability of expert opinion in rural areas Lack of outreach to the farmers Significant economic dependence on crops health Timely and accurate diagnosis of crops diseases is required with minimal dependence on experts 5
6 Proposed Model 6
7 Disease Diagnosis for Crops in Mianwali Region Crops: Wheat and cotton Number of diseases: 21 Number of symptoms in raw data: 50+ Number of instances: 160 Source: Field surveys (farmers) + online raw data + District Agriculture Department 7
8 Proposed Model 8
9 Fuzzification 9
10 Implication 10
11 Defuzzification 11
12 Disease Inference 12
13 Snapshot of Domain Data 13
14 Proposed Model 14
15 jfuzzylite: Library being used for end-product Free and open-source Fuzzy logic control library Programmed in Java Equipped with a wide variety of controllers, membership functions, aggregation and defuzzification methods 15
16 16
17 Results Test data of 100 instances collected from surveys 17
18 18
19 Outcome 19
20 MeraMaweshi Intelligent Disease Diagnosis System for Livestock in Rural Pakistan 20
21 The Problem About 40 millions population of the country relies upon livestock. This sub-sector contributes about 55% of agriculture and 12% of overall GDP. Pakistan is the 4 th largest producer of milk in the world. Domain experts are not easily available for helping the farmers as the ratio of available domain experts to number of cattle is very low. Remoteness and sparse rural populations are the main factors. Average price of a cow or buffalo is about Rs. 150,000. Their health is the key challenge faced by the farmers. The farmers of rural areas in Pakistan have weak access to veterinary experts to diagnose the disease of their cattle resulting into significant financial loss every year. 21
22 Findings of Literature Review Thin amount of work has been done in most of the systems Data is either small or unreal Most of the published work is at proposal stage and position papers have been published Most of the system have not been deployed for real users Some developed systems are commercial and a common farmer has to pay for subscription Validation of the proposed model is weak Information about the collected data and methodology is vague There is no work (to the best of our knowledge) contextualized for Pakistani farmer community 22
23 How it works! Identification of Research Problem Literature Survey Data Collection District Livestock Office Mianwali ( 7 vet hospitals in 3 tehsils) UVAS Pattoki Campus Intelligent Algorithm Design Implementation Deployment
24 Disease Coverage 89 clinical symptoms 33 diseases real data points UVAS Ravi Campus District Livestock Office Mianwali 24
25 Information Architecture 25
26 The Solution to the Problem 26
27 The Solution to the Problem 27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42 Comparison Matrix Best Options Solution 1 Solution 2 Solution 3 Our Solution Availability for Farmer No No No Yes Language Localization No No No Yes Confidence Level No No No Yes Medicine Suggestion No No No Yes Connect to Doctor No No No Yes Validation No No No Yes Solution 1: Cow Disease Diagnosis System in China Solution 2: Egyptian Bovine Clinical Knowledge for Newly Born Cows and Buffalos Solution 3: CattleToday: A web-based information system for cattle disease management 42
43 Validation Results
44 Availability (Urdu/English) Google Play Store: Mera Maweshi (Urdu/English) 44
45 Market / Opportunity THE TARGET MARKET Farmers, Veterinary Practitioners, Pharmaceutical Industry. TOTAL TARGET MARKET SIZE 7~8 Million Farmers, 5,000 Veterinary Practitioners, 50+ Pharmaceutical Companies 45
46 Marketing Plan Aggressive Outreach to Farmers Road Shows at Mela Maweshi Outreach to Rural Areas Aggressive outreach to Veterinary Practitioners Aggressive Outreach to Pharmaceutical Companies Social Media SMS Alerts Public seminars Conventional media 46
47 Way Forward Sophisticated HCI for illiterate farmers Handling complicated situations Involving more stakeholders 47
48 Funding Rs million research grant for Mera Maweshi Ministry of IT & Telecom Government of Pakistan 48
49 Collaborations 49
50 Conclusion ICT has positive impact in many disciplines of our daily life including healthcare, business, education, communication etc. Pakistan is an agricultural country with rich potential ICT has immense scope in bringing deeper impact on our agricultural economy Need: Inter-disciplinary and integrated teaching, research & collaborations amongst all stakeholders 50
51 Example of Netherlands Total cultivable area of Netherlands: ~2 million hectors Total cultivable area of Pakistan: ~20 million hectors Agri & food export of Netherlands in 2016: ~94 billion EUR Agri & food export of Pakistan in 2016 : ~3.3 billion EUR Unfortunately, our agriculture export declined by ~14% over the last two years Netherlands is the 2 nd largest exporter of agri & food products after US Germany, Brazil and China take 3 rd, 4 th and 5 th positions Why is it so? 51
52 Thank You for Your Attention 52
53 References [1] Economic Survey of Government of Pakistan ( ) [2] Food, A. O. of the United Nations, FAO Statistical Year Book 2013, FAO Publishers, [3] S. R. Sindhu, F. H. Shah, N. A. Khan, A. M. Jatt, S. Fazlani, A. Memon, Tax to GDP Ratio: Measures for improvement, Technical Report, Directorate of Training and Research (Inland Revenue Service) Lahore, Pakistan, [4] G. of Pakistan, Overview of The Economy, Technical Report, Ministry of Finance, Government of Pakistan, [5] M. El-Telbany, M. Warda, M. El-Borahy, Mining the classification rules for egyptian rice diseases, The International Arab Journal of Information Technology (2006). [6] M. Ilic, P. Spalevic, M. Veinovic, A. A. M. Ennaas, Data mining model for early fruit diseases detection, in: 23rd Telecommunications Forum (TELFOR), [7] J. Huang, Y. Yuan, W. Cui, Y. Zhan, Development of a data mining application for agriculture based on bayesian networks, in: Computer And Computing Technologies In Agriculture, volume 258, [8] H. Li, R. Ji, J. Zhang, X. Yuan, K. Hu, L. Qi, Web-based intelligent diagnosis system for cotton diseases control, International Federation for Information Processing (2011). 53